mirror of
https://github.com/nexu-io/open-design.git
synced 2026-06-01 03:14:35 +07:00
90 commits
| Author | SHA1 | Message | Date | |
|---|---|---|---|---|
Tom Huang
|
da2b007a43
|
feat(daemon): add DeepSeek TUI as a code agent adapter (#439)
* feat(daemon): add DeepSeek TUI as a code agent adapter
Register `deepseek` (with `deepseek-tui` cargo-only fallback) in
AGENT_DEFS via `deepseek exec --auto [--model X] <prompt>` and plain-text
streaming. Ships `deepseek-v4-pro` / `deepseek-v4-flash` as fallback
model hints; users can paste any other id (incl. NIM / Fireworks /
SGLang routes) via the custom-model input.
Web UI gets a DeepSeek-blue gradient icon, label/alias mapping, and
docs/agent-adapters.md §5.9 documents the auth state, prompt-as-argv
Windows size limit, and the upstream gap that prevents stdin delivery
today (clap declares `prompt: String` as a required positional).
Adds .deepseek/ to .gitignore alongside the other per-agent runtime
data dirs so first-launch trust files don't leak into git.
* fix(daemon): drop unsupported deepseek-tui fallback bin
The `deepseek` dispatcher owns `exec` / `--auto`; `deepseek-tui` is the
runtime companion it invokes. Listing `deepseek-tui` in fallbackBins
advertised availability for a host that only had the TUI binary, but
buildArgs still emitted `<resolved> exec --auto <prompt>` — which
deepseek-tui itself doesn't accept, so the first /api/chat run would
fail. Upstream documents both binaries as required (npm and cargo paths
install them together), so the fallback didn't correspond to a supported
install. Pin the absence in the agents test and update docs §5.9 + the
adapter table to match.
Generated-By: looper 0.4.0 (runner=fixer, agent=claude-code)
* fix(daemon): pre-flight DeepSeek TUI prompts against argv byte budget
DeepSeek's exec mode requires the prompt as a positional argv arg (no
`-` stdin sentinel upstream), so a fully composed OD prompt — system
text + history + skills + design-system content + the user message —
can blow Windows' ~32 KB CreateProcess limit (or Linux MAX_ARG_STRLEN
on extreme edges) and surface as a generic spawn failure instead of
a DeepSeek-specific, user-actionable message. The adapter now declares
`maxPromptArgBytes = 30_000` (leaves ~2.7 KB argv headroom for `exec
--auto --model <id>` and Windows quoting), and the /api/chat spawn
path checks the composed prompt against that budget before calling
`spawn`. Oversized prompts fail fast with `AGENT_PROMPT_TOO_LARGE`
and guidance to reduce skills/design context or pick an adapter with
stdin support.
Generated-By: looper 0.4.0 (runner=fixer, agent=claude-code)
* test(daemon): pin DeepSeek argv-budget guard with regression tests
The previous spawn-path guard inlined the byte-budget check in the
chat handler, so the only safety net for the DeepSeek argv-only
prompt-delivery shape was a static "the field exists" assertion —
nothing actually exercised the AGENT_PROMPT_TOO_LARGE path or the
short-prompt happy path. Extract the check into a pure
`checkPromptArgvBudget(def, composed)` helper in agents.ts, call it
from /api/chat before bin resolution (so the guard is order-
independent and fires regardless of whether the adapter binary is
on PATH in CI), and add a regression test that exercises both the
oversized-prompt branch (over the conservative under-Windows-
CreateProcess budget) and the short-prompt branch, plus a UTF-8
byte-vs-codepoint case and a stdin-adapter no-op case so the guard
can't silently regress or leak onto adapters that ship the prompt
over stdin.
Generated-By: looper 0.4.0 (runner=fixer, agent=claude-code)
* fix(daemon): pre-flight DeepSeek prompts against Windows .cmd-shim quoting
The first-pass argv-byte guard only inspects the raw composed prompt, so
on Windows an npm-installed `deepseek` resolves to a `.cmd` shim and the
spawn path then wraps the call in `cmd.exe /d /s /c "<inner>"` with
every embedded `"` doubled by `quoteWindowsCommandArg`. A quote-heavy
prompt (code blocks, JSON-shaped skill seeds) under the 30,000-byte
budget can therefore still expand past CreateProcess's 32_767-char
`lpCommandLine` cap and surface as a generic spawn ENAMETOOLONG instead
of the DeepSeek-named, actionable `AGENT_PROMPT_TOO_LARGE` the budget
guard was meant to provide. Add a second pure helper
`checkWindowsCmdShimCommandLineBudget(def, resolvedBin, args)` that
mirrors the platform layer's per-arg quoting and recomputes the would-be
command line length whenever the resolved binary is a `.cmd` / `.bat`
shim, and call it from `/api/chat` after `buildArgs` / `resolveAgentBin`
so the same SSE error fires before `spawn`. Pin the new path with a
quote-heavy regression (prompt is under the byte budget but doubles
past the kernel cap) plus no-op tests for non-`.cmd` resolutions, null
bin, and stdin-only adapters so the guard can't drift back.
Generated-By: looper 0.4.0 (runner=fixer, agent=claude-code)
* fix(daemon): extend DeepSeek argv guard to direct .exe Windows installs
The cmd-shim guard added in
|
||
emilneander
|
33c3b94b42
|
feat(daemon): add od mcp - expose Open Design as an MCP server (#399)
* feat(daemon): add `od mcp` subcommand for stdio MCP server
Lets a coding agent in a different repo (Claude Code, Cursor, Zed)
pull files from a locally-running OD project over the Model Context
Protocol — no export/import zip dance.
The MCP server is a thin stdio process that proxies read-only tool
calls to the daemon's existing HTTP API; no daemon-side changes
required. Exposes 8 tools:
list_projects, get_project,
list_files, get_file,
list_skills, get_skill,
list_design_systems, get_design_system
Wired exactly like `od media`: a hoisted flag set, a SUBCOMMAND_MAP
entry, a thin handler that resolves OD_DAEMON_URL and hands off to
src/mcp.ts. Tool dispatch is a switch over the tool name; each branch
fetches the matching daemon route and surfaces the response as MCP
text content. Binary mimes return a clear error pending phase-2
support.
Lifecycle gotcha worth flagging: Server.connect(transport) only
*starts* the stdio reader; the promise resolves immediately. Without
holding the function awaiting until transport/stdin close, cli.ts's
top-level process.exit(0) kills the server before the first request
arrives. The fix in src/mcp.ts holds until onclose / stdin EOF.
Wire-up example for a consuming repo:
{
"mcpServers": {
"open-design": {
"command": "od",
"args": ["mcp"],
"env": { "OD_DAEMON_URL": "http://127.0.0.1:7456" }
}
}
}
New dep: @modelcontextprotocol/sdk (MIT, official Anthropic SDK).
* feat(daemon): add MCP server instructions for zero-shot LLM context
Hand the consuming LLM a system-prompt-style overview of the OD
workflow so it picks the right tool without prompt-engineering on
the user's side. Mentions get_artifact and project-name resolution
ahead of their actual implementation; both ship in the same batch.
* feat(daemon): resolve MCP project args by UUID, name, or substring
Lets a consuming agent say `project: "recaptr"` instead of pasting a
UUID. Match order: exact id → exact name (case-insensitive) →
slug-normalized name (strips trailing " (N)", normalizes whitespace) →
substring (errors if multiple). UUID inputs short-circuit and never
hit the daemon.
* feat(daemon): surface entryFile and kind on MCP get_project response
Promote metadata.entryFile and metadata.kind to top-level fields so
consumers (including get_artifact in this branch) can find the entry
without digging through nested metadata blobs.
* feat(daemon): add MCP get_artifact tool for bundle retrieval
A design rarely lives in a single file. get_artifact pulls the entry
HTML/JSX plus every sibling it references (tokens CSS, JSX modules,
imported components) in one call, so a consuming agent doesn't need
to parse HTML and round-trip per file.
Three modes:
auto (default): BFS over relative <script src>, <link href>,
<img src>, <source/video src>, JSX import/from, CSS url(), with
depth cap 3 and a visited set. CDN, data:, mailto:, anchors, and
paths containing .. are skipped.
all: every textual file in the project (mirror of /archive
minus binaries).
shallow: just the entry file (same as get_file).
Output is a structured JSON blob with name/mime/size/content per
file and the project's manifest metadata at the top.
* feat(daemon): add /api/projects/:id/search route + MCP search_files
Server-side substring search across textual project files. Returns
file, 1-indexed line, and snippet, capped at 1000 matches. Exposed
through the MCP layer as search_files(project, query, pattern?, max?).
Treats the query as a literal substring (regex chars escaped) to
avoid catastrophic-backtracking attacks from LLM-supplied input.
Honors the project dir's existing path-safety guards via listFiles.
* feat(daemon): add since= filter to /files route + MCP list_files arg
Lets a consumer poll for "what's changed since I last looked" without
re-walking every file. Daemon-side: parse since= as ms, filter
listFiles output by mtime. MCP-side: forward as URL query.
* feat(daemon): expose skills and design systems as MCP resources
Catalog reads are stable reference material — they fit MCP's
resources surface (LLM-passive) better than tools (LLM-active).
Skills and design systems each become resources at
od://skills/<id>/SKILL.md and od://design-systems/<id>/DESIGN.md;
existing list_skills / get_skill / list_design_systems /
get_design_system tools remain as fallbacks for clients that don't
handle resources cleanly.
* fix(daemon): tighten MCP correctness in get_artifact and resources
Several silent-failure paths and minor footguns the first pass missed:
- get_artifact auto: the entry's own fetch now raises a clear
error instead of returning files: []. Previously a typo in
`entry:` looked like an empty project.
- get_artifact: invalid `include` value returns a clear error
listing the valid modes instead of silently behaving as auto.
- get_artifact all: includes binary files as metadata stubs to
match auto's behavior. Both modes are now strict supersets of
shallow.
- extractRelativeRefs: gate JS-only patterns (import/from/require/
dynamic-import) by file mime/extension so prose in markdown or
HTML doesn't generate spurious 404 round-trips on words like
"imported from 'X'".
- extractRelativeRefs: cover <iframe>, <audio>, srcset, and
CSS @import — common in real OD output.
- resources/list descriptions are collapsed to a single line
(newlines + repeated whitespace -> one space) so MCP UIs that
don't normalize whitespace render cleanly.
- fetchProjectFile: 0-byte binary files no longer report size: null
due to falsy short-circuit on Number(content-length).
* perf(daemon): cache MCP project list for 5s in resolveProjectId
A typical agent session calls list_files/get_file/get_artifact several
times in a row, each with a project name argument. Each previously
re-fetched /api/projects. Cache the list in module scope with a 5s
TTL so back-to-back lookups are local; renames in the OD UI still
propagate within a few seconds.
* feat(daemon): MCP UX polish — tool order, annotations, get_artifact maxBytes
Three changes well-behaved MCP clients pick up automatically:
- Tool ordering. list_projects + get_artifact are now first; LLMs
that weight earlier entries surface the bundle path before
per-file fetching. Catalog tools (list_skills, get_skill,
list_design_systems, get_design_system) sit at the bottom; they
are also exposed as MCP resources.
- readOnlyHint / idempotentHint / openWorldHint annotations on
every tool so clients can skip confirmation prompts on safe
tools and let the LLM know re-running is fine. Per-tool `title`
annotations give clients a friendlier display name than the
snake_case tool id.
- get_artifact gains a `maxBytes` arg (default 1.5MB). Once the
accumulated textual content crosses the cap, remaining files
are dropped and `truncated: true` is set on the bundle so the
consumer knows to use list_files / get_file for the rest.
* feat(daemon): expose user's active OD project/file via MCP
The "what file are you on?" round-trip the agent had to do every
session is now answered automatically. Three pieces:
- Daemon: in-memory active-context slot with 5-minute TTL.
POST /api/active sets {projectId, fileName}; GET /api/active
returns the current value enriched with projectName, or
{active:false} when the slot is empty/stale. Cleared on
daemon restart.
- Web: a small useEffect in App.tsx posts the active project +
file to the daemon on every route change. Best-effort fire-
and-forget; a missing daemon doesn't surface an error.
- MCP: get_active_context tool (no args) and a matching MCP
resource at od://focus/active. The tool is listed second,
right after list_projects, so an LLM picks it up before
asking for ids. Server instructions tell the model to call
it FIRST when the user says "this file" / "the design I have
open" / "what I'm looking at."
End to end: user opens a project in OD, agent in another repo
calls get_active_context() → gets {projectName: "recaptr",
fileName: "recaptr-onboarding-4.html"}, then immediately calls
get_artifact(project: "recaptr") with no further user input.
* feat(daemon): make MCP project arg optional, fall back to active OD context
get_artifact, get_project, get_file, search_files, and list_files now
accept project as optional. When omitted, the MCP resolves project
from /api/active so an agent in another repo can call
search_files({ query: "Polaroid" })
without first asking the user "which project?". get_file and
get_artifact also default their path/entry to the active file, so
get_file({}) returns whatever the user is currently looking at.
The implicit path stamps `usedActiveContext` on JSON responses (or a
separate `[od:active-context …]` content block on get_file) so the
agent can see exactly which project/file got chosen. Explicit
project args pass through with zero added overhead.
Cuts the common case from two MCP round trips
(get_active_context → search_files) to one. Server instructions and
get_active_context's own description are updated to point at the
new default.
* fix(daemon): require same-origin for /api/active POST and GET
The active-context endpoint was added without isLocalSameOrigin
guard. Since the daemon binds 0.0.0.0 by default, a LAN peer could
GET it to learn what file the user has open, or POST it to redirect
the MCP fallback to a project of their choice. Same-origin only is
the right scope: the web app proxies its requests through Next.js
on the daemon port, and the MCP runs over loopback in-process, so
both legitimate callers pass.
Pattern matches the existing /api/app-config etc. guards.
* feat(daemon): add /api/mcp/install-info for cross-platform install snippets
The Settings -> MCP server panel needs absolute paths to node and
the daemon's built cli.js so it can render snippets that work on a
fresh source clone (where `od` is not on PATH) and dodge the
/usr/bin/od octal-dump tool that ships on macOS/Linux and would
otherwise shadow ours.
Endpoint returns:
- command: process.execPath (the node binary running the daemon)
- args: [<absolute path to dist/cli.js>, "mcp"]
- daemonUrl: http://127.0.0.1:<port>
- platform: process.platform (so the panel can localize ~/.cursor
vs %USERPROFILE%\.cursor and Cmd vs Ctrl shortcuts)
- cliExists / nodeExists: existsSync checks on both binaries
- buildHint: human-readable build/reinstall instructions when
either path is missing
isLocalSameOrigin guard same as /api/active. Cached for 5s because
the panel may re-fetch on every open and the paths cannot change
without a daemon restart.
Test file covers the happy path, cross-origin rejection, two
allowed-Origin variants, and the cache by counting fresh resolves
across rapid calls. 5/5 pass.
* refactor(daemon): tighten MCP surface, trim descriptions, polish copy
Three intertwined cleanups that all live in mcp.ts + cli.ts:
1. Drop catalog tools from MCP. list_skills / get_skill /
list_design_systems / get_design_system are removed. The audience
is a coding agent in a separate repo consuming Open Design's
output; it cannot run skills (those are recipes Open Design uses
to generate) and design-system DESIGN.md is reference material
that already ships as an MCP resource. Keeping the catalog as
tools cost ~350 token-overhead per turn for capabilities the
agent could not act on. Tool count: 11 -> 7.
2. Trim tool descriptions. The active-context fallback explanation
was repeated in 5 separate tool descriptions; hoisted into
PROJECT_ARG and explained once in the server `instructions`
block instead. Saves ~150-200 tokens per tools/list response.
3. User-facing branding pass. Tool titles, tool descriptions,
resource names, error messages, comments, and `od mcp --help`
now consistently use "Open Design" rather than "OD". Internal
abbreviation `OD` is retained only inside the server
instructions block where it is introduced inline as "Open Design
(OD)" for compactness across multi-paragraph guidance.
Em dashes replaced with hyphens throughout, per project style.
* feat(web): add MCP server install panel in Settings
New "MCP server" section in the Settings dialog, surfacing
copy-paste install snippets for the major MCP-compatible coding
agents (Claude Code, Cursor, VS Code, Antigravity, Zed, Windsurf).
Highlights:
- In-brand custom dropdown (reuses the existing .ds-picker
pattern from the design-system / prompt-template pickers, click
outside / Escape to close, chevron animates) instead of a
native <select>.
- Per-client snippet that uses absolute paths to node + cli.js
fetched from /api/mcp/install-info on mount, so it works even
when `od` is not on PATH.
- Cursor gets a one-click "Install in Cursor" deeplink
(cursor://anysphere.cursor-deeplink/mcp/install) that pops an
approval dialog and writes the config for the user. UTF-8-safe
base64 so paths with accented characters do not throw.
- Per-OS path hints (~/.cursor on POSIX, %USERPROFILE%\.cursor
on Windows) and keyboard shortcuts (Cmd vs Ctrl).
- Build-required warning card when cli.js or the node binary
does not exist on disk; deeplink button disables in that state.
- Prominent "restart your client to pick up the new server"
callout below the snippet, with per-client guidance.
- Capability list ("what your agent can do") instead of a tool-
name dump, so non-developer designers can also tell what is
possible without reading MCP docs.
README adds a short "Use Open Design from your coding agent"
section that points at the panel and summarizes the per-client
flow (one-click for Cursor, JSON merge elsewhere). Read-only by
design; the daemon must be running locally.
* docs(readme): align MCP server section with the Settings panel
The "Use Open Design from your coding agent" section had drifted
from what the panel actually emits and lists.
- Add Antigravity to the supported-client list (previously missing).
- Drop the "(GitHub Copilot)" parenthetical from VS Code so the
label matches the panel.
- Fix the Claude Code line: we no longer emit a single
`claude mcp add ...` shell command. The snippet is JSON; the
panel additionally suggests `claude mcp add-json` as the safer
way to apply it instead of hand-editing ~/.claude.json.
- Swap the "find the Polaroid section" example for two more
universal phrases ("build this in my app", "match these
styles") that match what the panel surfaces.
- Add a one-line "restart or reload your client after install"
note - this was prominent in the panel and absent from the
README.
- Trim the /usr/bin/od octal-dump aside; it was technical detail
that did not earn its space at the README intro level.
* feat(web): add Codex CLI to the MCP server install panel
Codex is a first-class supported coding agent (listed alongside
Claude Code, Cursor, etc. in the README's PATH-detected agent
table) but the install panel was missing it.
Codex stores MCP server config at ~/.codex/config.toml (TOML, not
JSON) under an `[mcp_servers.<name>]` table, and the same file is
shared between the Codex CLI and the Codex IDE extension - so one
install covers both. Added a 7th client entry that emits the right
TOML snippet, expanded the snippet-lang union to include 'toml'
(behaves like 'json' for whitespace handling, just a different
syntax-highlight hint).
For our minimal payload (just command + args), JSON.stringify
happens to produce valid TOML literal values since TOML basic
strings use the same double-quote escape rules as JSON, and TOML
inline arrays match JSON array syntax. No new TOML serializer
needed.
README updated to list Codex among the supported clients.
Schema verified against https://developers.openai.com/codex/mcp.
* fix(daemon): accept any loopback origin in same-origin guard
The previous port-pinned check required the request's Origin to match
either the daemon's own port or OD_WEB_PORT. tools-dev does not pass
OD_WEB_PORT to the daemon process, so any browser POST to /api/active
proxied through the dev web (port 17573 etc.) was rejected with 403,
and get_active_context always returned {active: false}.
Relax to a loopback-prefix match: any http://127.0.0.1:*,
http://localhost:*, or http://[::1]:* origin passes regardless of
port. Cross-origin (https://evil.com) is still rejected. The
trade-off is that another local web app on a different loopback port
could now CSRF the daemon; same-origin checks are inherently a CSRF
defense, not a network ACL.
* fix(web): make Claude Code MCP snippet a real copyable one-liner
claude mcp add-json open-design '<json>' takes only the inner
server-config object, not the full {"mcpServers": ...} wrapper, and
rejected the wrapped shape with "Invalid configuration: : Invalid
input". Pass only the inner config, and inline the JSON into the
command itself so the snippet is a real one-liner the user can copy
and paste, no template substitution.
* test(daemon): drop loopback-prefix assertions superseded by upstream origin policy
The two proxy-flow allow tests were added in
|
||
Daniel Duma, PhD
|
cfd359e05a
|
[codex] Fix Gemini CLI trust handling (#352)
* Fix Gemini CLI trust handling * Preserve agent spawn env filtering |
||
Bryan A
|
d637297313
|
feat(preview): live-reload iframes when project files change on disk (#409)
* feat(preview): live-reload iframes when project files change on disk Add a chokidar-backed file watcher per active project on the daemon, surface changes via an SSE endpoint at /api/projects/:id/events, and consume them in the web app to bump the file list. The new mtime then propagates to the FileViewer iframe through PR #384's ?v=${mtime} cache-bust, reloading the preview automatically — no manual refresh click. Daemon: - New apps/daemon/src/project-watchers.ts: refcounted per-project watcher registry. First subscribe lazy-creates a chokidar watcher; last unsubscribe closes it. Ignores .git, node_modules, .od, debug, .DS_Store. Returns a ready promise so callers can await initial scan. - New endpoint GET /api/projects/:id/events using the existing createSseResponse helper. Sends one ready event after chokidar binds, then one file-changed event per add/change/unlink. - Adds chokidar ^5.0.0 dependency. Web: - New apps/web/src/providers/project-events.ts exposing createProjectEventsConnection (pure, testable) and useProjectFileEvents hook. EventSource with exponential backoff (1s -> 30s cap), reset on a successful ready event. - ProjectView.tsx subscribes when daemonLive && project.id, and on each event bumps the existing filesRefresh signal — no FileViewer changes needed because PR #384 already URL-loads with mtime cache-bust. Tests: - 6 new daemon unit + integration tests (refcounting, real chokidar add/change/unlink, ignore patterns). - 8 new web hook unit tests (URL encoding, payload parsing, malformed payload tolerance, exponential backoff, backoff reset on ready, close cancels reconnects, no-op when EventSource missing). Closes #370 Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * fix(preview): test ignore patterns relative to watch root The ignore predicate matched against absolute paths, so segments in the watch root's ancestors (e.g. the daemon's own .od/ runtime dir, which contains every project) silenced every event. In production this meant zero file-changed events ever fired — every file inside a project sat under .od/projects/<id>/, and .od matched the ignore. Tests passed because mkdtemp puts test roots in /tmp/od-watchers-XXX/, which has no .od ancestor. Fix: compute the path relative to the watch root, then test segments. Add a regression test that reproduces the production layout (.od/projects/<id>/...) and asserts events still fire. Also folds in a small consolidation of the SSE route handler from the prior commit on this branch (single route, surfaces sub.ready before emitting `ready`, propagates err.message in the error path). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * fix(project-view): only auto-open files that exist in the project Agent Write/Edit tool results unconditionally called requestOpenFile on the basename of the edited path, which created permanent placeholder tabs ("Open a file from Design Files.") whenever the agent edited a file outside the project's working directory (e.g. an upstream repo source file). Add decideAutoOpenAfterWrite() — a pure helper that gates the auto-open on the file actually appearing in the refreshed project file list. Same nextFiles-from-then() pattern already used at ProjectView.tsx:968. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * fix(preview): chokidar resilience + auto-open path-suffix matching Address PR #409 review feedback in one bundle: - chokidar error listener (codex P1): FSWatcher is an EventEmitter; without an error handler, transient FS faults (ENOSPC, EPERM, EMFILE) surfaced as unhandled exceptions and could crash the daemon. Watcher now logs in dev mode and continues; refcount cleanup unaffected. - followSymlinks: false (mrcfps): keep the watcher's resource boundary aligned with the project boundary so a symlink inside the project cannot traverse externally. Real-chokidar regression test included. - decideAutoOpenAfterWrite path-suffix matching (mrcfps): pass the agent's full file_path through (not just the basename); resolve via path-suffix match against project file paths, with single-unambiguous basename fallback only when filePath has no slash. Fixes the same-basename collision case where an external Write to App.jsx could open a project's prototype/App.jsx. - Dev-mode logs (lefarcen P3): warn when the broadcast subscriber loop or the SSE payload parser swallows an error, so subscriber bugs and payload-shape regressions don't go silent during testing. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> --------- Co-authored-by: DevForgeAI CI/CD Engineer <devforge-ai@development.ai> Co-authored-by: Claude Opus 4.7 (1M context) <noreply@anthropic.com> |
||
|
Tom Huang
|
aefba56a3f
|
feat(skills): open-design-landing rename, kami skills, landing OG (#428)
* feat(skills): open-design-landing rename, kami skills, landing OG - Rename editorial-collage skills to open-design-landing and -deck; refresh examples and compose script layout - Add kami-deck and kami-landing skills with HTML examples - Landing page: og.astro, index wiring, and style tweaks; package.json bump - Web i18n: German and Russian copy for renamed and new skills - Daemon test: update skill-asset-rewrite expectations for new paths - Design systems: README and atelier-zero doc touch-ups - Cross-skill SKILL.md reference updates Co-authored-by: Cursor <cursoragent@cursor.com> * docs(landing-page): document version-slot invariant and deprecation timeline Address P3 review notes on PR #428: - Note the `data-github-version` wrapper invariant (version string only) near the canonical URL block in `app/page.tsx`. - Expand the `formatVersion` helper comment in `app/pages/index.astro` with concrete `release.name` / `tag_name` example shapes for each branch of the regex fallback. - Tighten the `EditorialCollageDeckInputs` deprecation in `skills/open-design-landing-deck/schema.ts` to a specific removal version (v0.4.0) and add a "Migrating from editorial-collage-deck" section to the skill README. Generated-By: looper 0.4.0 (runner=fixer, agent=claude-code) * docs(landing-page, skills): clarify version slot script and rename migrations - Describe GitHub version slots as driven by the inline enhancement script, not React hydration. - Add editorial-collage → open-design-landing migration notes; fix README link copy (Astro static landing app). - Extend deck README migration table with shared asset path renames. Co-authored-by: Cursor <cursoragent@cursor.com> * fix(daemon): alias deprecated editorial-collage skill ids The PR renames the editorial-collage / editorial-collage-deck skills to open-design-landing / open-design-landing-deck, but the daemon persists exact skill_id strings on projects and resolves them via listSkills().find((s) => s.id === storedId). After the rename, any project saved against an old id silently composes without the intended skill prompt because the listing no longer exposes that id. Add a SKILL_ID_ALIASES map in skills.ts plus a findSkillById() helper that rewrites deprecated ids to their current canonical form, then route every server-side lookup (skill detail, example HTML, asset proxy, system-prompt composer) through it. Cover the alias map, the resolver, and end-to-end resolution against a temp skills directory with a regression test. Generated-By: looper 0.4.0 (runner=fixer, agent=claude-code) * fix(kami-deck): route host od:slide messages through local go() The host bridge classifies kami-deck as class-driven because go() toggles .slide.active, but the visible slide is moved by deck.style.transform which the bridge cannot drive. Listen for od:slide messages and dispatch them through the local go() so toolbar next/prev and initialSlideIndex restore actually shift the deck. Generated-By: looper 0.4.0 (runner=fixer, agent=claude-code) * fix(kami-deck): sync deck transform with host-driven .active changes The previous fix added a local od:slide listener but the host bridge in apps/web/src/runtime/srcdoc.ts also listens for the same message and calls setActive() (toggles .slide.active) without driving the deck transform. Both listeners fired, the bridge re-read the just-toggled active class, and overshot by one — and the bridge's restoreInitialSlide path could move .active without a message at all, leaving the deck on the original transform. Stop the bridge from double-handling by calling stopImmediatePropagation in the local listener (registered first because the bridge script is appended to </body>), and add a MutationObserver that pulls the deck transform onto whichever slide currently carries .active so the bridge's direct setActive calls (notably the initial-slide restore) move the deck too. Generated-By: looper 0.4.0 (runner=fixer, agent=claude-code) * fix(i18n): align French content with renamed/new skills PR #434 (French localization) merged into main with French copy for the old editorial-collage / editorial-collage-deck skill ids; this branch renamed those to open-design-landing / open-design-landing-deck and added kami-deck and kami-landing. Update content.fr.ts to track the rename and add French copy for the new kami skills so the LOCALIZED_CONTENT_IDS coverage test passes once main is merged. Generated-By: looper 0.4.0 (runner=fixer, agent=claude-code) * fix(open-design-landing-deck): sync deck transform with host-driven .active changes Apply the same fix that landed in skills/kami-deck/example.html (commits |
||
monshunter
|
bf533c3d72
|
fix(web): split execution mode tabs and align active chip visuals (#418)
* Refine settings API configuration UX * Fix PR validation issues * Address settings review edge cases * Address BYOK settings review feedback * Fix BYOK provider labels and model display |
||
Xinmin Zeng
|
76b41ffef4
|
fix(daemon): expose skill resources via cwd-relative aliases (#435)
* fix(daemon): expose skill resources via cwd-relative aliases (#430) Skill bodies referenced repository-absolute paths in their preamble and relied on every agent CLI honouring `--add-dir` (or its equivalent) to open them. Most agents (codex, gemini, opencode, cursor-agent, qwen, pi, hermes, kimi) have no such flag, and Claude Code's directory-access policy can still block these reads under specific permission/trust configurations even when `--add-dir` is set, so SKILL side files fail to load mid-turn. Stage `.od-skills` and `.od-design-systems` directory links inside the project cwd before spawning the agent, and rewrite the skill preamble to address resources via the cwd-relative alias path. The alias resolves through a directory junction (Windows) or symbolic link (POSIX), so resources are reachable from inside every agent's working directory without needing an extra-allowed-directories flag. `extraAllowedDirs` is kept as a belt-and-suspenders fallback for Claude/Copilot — if the alias cannot be created (e.g. the user has a pre-existing real entry under one of the reserved names), `--add-dir` still covers the absolute paths. `ensureCwdAliases` is idempotent, repoints when the target moves, and refuses to overwrite a non-symlink entry under the reserved name to avoid clobbering user data. Closes #430. * fix(daemon): address review feedback — copy active skill, dual preamble paths Round 2 of PR #435 addressing the unresolved P1/P2/P3 review threads from @chatgpt-codex-connector, @lefarcen and @mrcfps. - P1 writable-symlink. Replace the directory link with a per-project `fs.cp` of the active skill into `<cwd>/.od-skills/<folder>/`. Writes through the staged copy no longer mutate the shipped repo resource. A regression test pins this down so a future "optimisation" that re-introduces a symlink would fail loud. - P1 absolute fallback. Rewrite the skill preamble to advertise both the cwd-relative alias path (primary) and the absolute path (fallback). When `stageActiveSkill` is skipped (no project cwd) or errors, Claude/Copilot still reach the resource via the absolute path covered by `--add-dir`; other agents that fall back to `PROJECT_ROOT` as cwd reach it natively because the absolute path is in-cwd there. - P2 symlinked source root. Use `stat()` instead of `lstat()` on the source so a symlinked `SKILLS_DIR` or skill folder is followed rather than skipped. - P2 stale ordering. `ensureCwdAliases` is gone; staging now happens independently of prompt composition and never silently diverges from what the preamble advertises. - P3 unsafe spec name. Validate `folderName` against an explicit segment policy (no path separators, no dot segments, no absolute, no null bytes) so a malformed caller cannot escape the alias root. - P3 strong types. Drop the `@ts-nocheck` pragma on `cwd-aliases.ts` and type the public API (`SkillStagingResult`, `SkillStagingLogger`). Also drops the `.od-design-systems` alias entirely. Design-system bodies are read by the daemon and folded into the system prompt directly, so an agent never has to open them through the filesystem; exposing that tree as cwd-visible was unnecessary surface area. Test additions: 17 cases in `cwd-aliases.test.ts` covering the copy semantics, the write-barrier regression, symlinked-source handling, legacy-symlink upgrade, alias-root collision refusal, cwd=null guard, and a parameterised set of unsafe folder names. * fix(daemon): unify effectiveCwd and use junctions on Windows test fixtures Round 3 of PR #435 addressing two non-blocking follow-ups from @mrcfps: - server.ts: collapse the chat-handler `cwd` resolution into a single `effectiveCwd = cwd ?? PROJECT_ROOT` and feed it consistently to `buildArgs`, `spawn`, and the ACP/Pi RPC session adapters. Previously `spawn(... { cwd: cwd || undefined })` let no-project runs inherit whatever working directory the daemon process was started from (packaged daemons / service launches do not start in the workspace root), which made the "absolute path is in-cwd" guarantee in the new skill preamble untrue for that exact case the comment claimed to cover. With `effectiveCwd`, agents in no-project mode land in `PROJECT_ROOT` and the absolute fallback path emitted by the preamble is genuinely a relative-to-cwd path. - cwd-aliases.test.ts: replace `symlinkSync(..., 'dir')` fixtures with a platform-aware `dirLinkType` (`'junction'` on Windows, `'dir'` elsewhere). Directory symlinks on Windows require SeCreateSymbolicLinkPrivilege / Developer Mode, which is not enabled on most CI images, so the previous form would fail the daemon test suite on Windows runners even though the production code (which uses `fs.cp`) does not need symlinks at all. |
||
|
Tom Huang
|
6c2a8ba09f
|
feat(editorial-collage): introduce Atelier Zero style landing page as… (#366)
* feat(editorial-collage): introduce Atelier Zero style landing page assets and documentation - Added new design system for Atelier Zero, including a detailed `DESIGN.md` file. - Created an `editorial-collage` skill with associated assets for a magazine-grade landing page. - Included example HTML and image assets for various sections (hero, about, capabilities, etc.). - Updated README files to guide usage and customization of the new skill and design system. - Introduced a new image generation prompt pack for consistent visual style across the landing page. * fix(i18n): cover atelier-zero design system and editorial-collage skill in German content Generated-By: looper 0.4.0 (runner=fixer, agent=claude-code) * fix(editorial-collage): align manifest with shipped assets and address PR review - Update image-manifest.json widths/heights/ratios to match the actual PNGs on disk: hero/about/cap/testimonial/cta = 1024x1024 (1:1), method-1..4 = 816x816 (1:1), lab-1..5 and work-1..2 = 768x1024 (3:4). Mirror the new dimensions in imagegen-prompts.md headings and in README.md. - Mark testimonial.png as rekey_on_brand_change so the manifest agrees with SKILL.md's "regenerate at minimum testimonial.png" guidance, and add work-1/work-2 to the rekey list in SKILL.md and README.md. - Add a Hero (I.) sec-rule and renumber every following section II..VIII in example.html so the eight sections walk sequentially I -> VIII and the page-of-008 counter starts at 001. - Delete editorial-artifact-system/ (16 duplicate PNGs + index.html + skills.md draft) — the canonical version is skills/editorial-collage/ and the duplicate had no consumer references. - DESIGN.md: spell out which dimensions of each magazine reference (Monocle/Apartamento/IDEA), document the rationale for single-accent vs multi-accent, and extend the anti-pattern list with AI-image-gen artifacts the system explicitly rejects. - SKILL.md: add italic_words validation guidance (trim, cap at 4, verb->noun rewrite, punctuation strip) and replace the broken-image fallback with an inline SVG placeholder sized to the slot's manifest aspect ratio. Generated-By: looper 0.4.0 (runner=fixer, agent=claude-code) * fix(daemon): serve skill example assets via stable API route Skill example HTML such as `skills/editorial-collage/example.html` references shipped images via `./assets/*.png`. The web app loads the example into a sandboxed iframe via `srcdoc`, where relative URLs resolve against `about:srcdoc` and the PNGs render as broken images in the Examples preview. Add a `GET /api/skills/:id/assets/*` route that serves files under the skill's `assets/` directory with path-traversal guards, and rewrite `src='./assets/<file>'` / `href='./assets/<file>'` in the example response to point at that route. The disk preview keeps working because the on-disk files are unchanged. Generated-By: looper 0.4.0 (runner=fixer, agent=claude-code) * feat(landing-page): add new static Next.js 16 site for Open Design marketing - Introduced a new landing page application using Next.js 16, featuring a static export setup. - Added essential files including `package.json`, `next.config.ts`, and TypeScript configuration. - Implemented global styles in `globals.css` to match the Atelier Zero design system. - Created a detailed `AGENTS.md` for module-level boundaries and purpose. - Included various image assets for the landing page, ensuring a visually cohesive experience. - Established a root layout and main page structure to support the marketing content. * style(landing-page): enhance topbar layout and improve responsiveness - Added nowrap styling to topbar elements to prevent text overflow. - Introduced media query to hide mid text in the topbar for screen widths between 1200px and 1280px. - Updated layout.tsx to suppress hydration warnings for better rendering consistency. - Removed redundant "Compiled by Open Design" text from the page component. * feat(landing-page): implement scroll-reveal animations for enhanced user experience - Added a new `RevealRoot` component to manage scroll-triggered reveal animations. - Updated `globals.css` with styles for elements using the `data-reveal` attribute, including opacity, translation, and scaling effects. - Modified `layout.tsx` to include the `RevealRoot` component for managing animations. - Enhanced `page.tsx` by adding `data-reveal` attributes to various elements for staggered reveal effects. - Implemented reduced motion support to ensure accessibility for users with motion sensitivity. * fix(landing-page): update import paths and enhance link styles - Changed the import path in `next-env.d.ts` to reference the correct routes type definition. - Enhanced `globals.css` with new styles for topbar links, work cards, and partner elements, improving hover effects and transitions. - Updated `page.tsx` to include canonical project URLs and made various links point to these URLs for better navigation and accessibility. * feat(landing-page): implement headroom-style sticky header with live GitHub star count - Introduced a new `Header` component to manage sticky navigation behavior on scroll, enhancing user experience. - Updated `globals.css` to style the sticky header, including transitions and visibility toggling based on scroll direction. - Modified `page.tsx` to replace the static header with the new `Header` component, which fetches and displays the live GitHub star count. - Ensured accessibility by providing a fallback for users who prefer reduced motion. * feat(landing-page): enhance editorial landing page with global ticker and new styles - Updated `next-env.d.ts` to reference the correct routes type definition for development. - Enhanced `globals.css` with new styles for the global ticker, including responsive design and improved overflow handling. - Introduced a new `WIRE_CITIES` and `WIRE_CONTRIBS` data structure in `page.tsx` to display a counter-scrolling marquee of cities and contributors. - Added a ghost button style for the navigation call-to-action in the header. - Updated various sections in `page.tsx` to integrate the new ticker and improve overall layout and accessibility. * refactor(landing-page): update paper texture overlay and remove multica-ai link - Enhanced comments in `globals.css` to clarify the purpose and behavior of the paper texture overlay. - Adjusted z-index of the overlay to ensure proper layering with other elements. - Removed the `multica-ai` partner link from `page.tsx` to streamline the partner section. * feat(landing-page): implement dynamic contributor marquee with GitHub integration - Added a new `Wire` component to display a counter-scrolling marquee of cities and contributors. - The contributor list is fetched live from the GitHub API, ensuring up-to-date information. - Updated `page.tsx` to integrate the `Wire` component, replacing the static contributor list with dynamic content. - Enhanced comments for clarity regarding the functionality and purpose of the global wire. * fix(i18n): add German display copy for editorial-collage-deck skill The Validate workspace test asserts that GERMAN_CONTENT_IDS.skills covers every curated skill on disk; the new editorial-collage-deck skill was missing from DE_SKILL_COPY, causing src/i18n/content.test.ts to fail. Generated-By: looper 0.4.0 (runner=fixer, agent=claude-code) * feat(landing-page): migrate marketing site to Astro * perf(landing-page): remove React client runtime * perf(landing-page): serve images from Cloudflare resizing * fix(pr): address landing page review feedback --------- Co-authored-by: mrcfps <mrc@powerformer.com> |
||
Sebastian Westberg
|
eb9693d56c
|
fix(daemon): strip ANTHROPIC_API_KEY when spawning Claude Code (#400)
* fix(daemon): strip ANTHROPIC_API_KEY when spawning Claude Code Claude Code prefers ANTHROPIC_API_KEY over the user's logged-in subscription auth when both are present, silently billing API usage even though the daemon's "Local CLI" mode implies the user wants their `claude login` credentials (Pro/Max plan) to be used. The daemon inherited the launching shell's env, so anyone who ever exported ANTHROPIC_API_KEY for SDK work or scripts hit this. Strip ANTHROPIC_API_KEY from the env passed to spawn for the claude adapter only, so Claude Code's own auth resolution applies. Fixes #398 * fix(daemon): strip ANTHROPIC_API_KEY case-insensitively, extract helper, add tests Address review feedback on #400: - Extract spawnEnvForAgent helper in agents.ts so the strip is unit- testable and lives next to the other agent-spawn primitives. - Strip ANTHROPIC_API_KEY case-insensitively. Windows env-var names are case-insensitive at the kernel level; spreading process.env into a plain object loses Node's case-insensitive accessor, so a mixed-case Anthropic_Api_Key would survive `delete env.ANTHROPIC_API_KEY` and still reach Claude Code via the spawned env block on Windows. - Add four tests covering: uppercase strip for claude, case-insensitive strip for claude, preservation for non-claude adapters, and input immutability. --------- Co-authored-by: Sebastian Westberg <sebastianwestberg@Sebastians-MacBook-Pro.local> |
||
nikshh
|
3b3275ea3d
|
fix(daemon): support nested paths in project file serve route (#401)
The /api/projects/:id/files/:name route used a named Express param that does not match path separators, so requests for files stored in subdirectories (e.g. assets/logo-mark.svg, fonts/Aeroport.woff2) returned 404. Switch to a wildcard route and read req.params[0] so that any depth of nesting is handled correctly. This fixes missing media and fonts when opening an HTML project file as a standalone preview. |
||
lefarcen
|
179b23d7e7
|
Modernize multi-provider API proxy routing
Adds provider-specific API proxy routes for Anthropic, OpenAI-compatible, Azure OpenAI, and Google Gemini; normalizes provider SSE streams; updates web provider clients/settings/docs/tests; and forwards Gemini token limits. |
||
Vitaliy VVS
|
64609e6a2a
|
Refactor RUNTIME_DATA_DIR resolution logic (#391)
* Refactor RUNTIME_DATA_DIR resolution logic Set OD_DATA_DIR=$HOME/.open-design #390 * Refactor resolveDataDir to accept projectRoot parameter Add validation if the resolved path is not writable. |
||
Jason
|
339a57c65c
|
fix(security): bind daemon to localhost by default, add origin validation (#365)
* fix(security): bind daemon to localhost by default, add origin validation middleware The daemon's Express server previously defaulted to binding 0.0.0.0, exposing all 30+ API endpoints to the local network with zero authentication. This included endpoints that spawn CLI agents, write project files, and manage API keys. Changes: - Default bind address changed from 0.0.0.0 to 127.0.0.1 in both server.ts and cli.ts. Users who need network access can still set OD_BIND_HOST=0.0.0.0 explicitly. - Added origin validation middleware on /api/* routes: requests with an Origin header are only allowed from localhost/127.0.0.1/[::1] on the daemon's own port. Non-browser clients (no Origin header) are unaffected. - 7 unit tests for the origin validation logic. - Existing tests: 327/327 passing. * fix: address PR review — fail-closed, OD_WEB_PORT, HTTPS, Origin: null Address feedback from @lefarcen, @mrcfps, and Codex review: P1: Fail-closed when resolvedPort is not yet set (returns 403 instead of passing through). P2: Include OD_WEB_PORT in allowed origins for split-port proxy setups (web port ≠ daemon port). Add HTTPS variants. P3: Exempt Origin: null for sandboxed iframe preview fetches (/api/projects/:id/raw/*) so artifact previews keep working. P4: Update help text from 0.0.0.0 to 127.0.0.1. Test coverage expanded to 13 tests including: - OD_WEB_PORT split-port proxy scenario - Origin: null iframe preview - HTTPS origin variants - Fail-closed before port resolution - Cross-origin rejection with OD_WEB_PORT set All 333 tests passing, TypeScript clean. * fix: scope Origin: null bypass to raw-file previews only, support non-loopback bind host Two issues found in second review round: 1. Origin: null was a global bypass for ALL /api routes, allowing sandboxed iframes to reach state-changing endpoints (project create/delete, agent runs). Now only GET requests to /projects/:id/raw/* pass through with Origin: null — the original intent for iframe file previews. POST/DELETE with Origin: null are rejected. 2. Allowed origins only included loopback addresses. When daemon binds to a non-loopback address (--host for Tailscale, LAN, or 0.0.0.0), browser requests from that address would get 403. Now the bound host is included in allowed origins alongside loopback, keeping the documented network-access escape hatch working for browser clients. Added negative tests for Origin: null on POST/DELETE/non-raw GET, and positive tests for non-loopback bind host scenarios. All 339 tests passing. * fix: centralize origin policy, add spritesheet to Origin: null allowlist Addresses @mrcfps's third review round: 1. isLocalSameOrigin() now uses the same policy as the global origin middleware: HTTPS + HTTP, OD_WEB_PORT, and the explicit bind host (OD_BIND_HOST). Previously it only accepted HTTP on loopback hosts, so requests from https://127.0.0.1 or a Tailscale address could pass the global guard but 403 on per-route checks. 2. /api/codex-pets/:id/spritesheet is a read-only route that sets Access-Control-Allow-Origin: null for canvas drawing by sandboxed iframes. Added to the Origin: null allowlist so the middleware doesn't block it before the route handler runs. 3. buildAllowedOrigins() extracted as a closure so both the middleware and isLocalSameOrigin() share identical logic. 340/340 tests passing. * test: remove dead fail-closed test with unused 5th argument The 'fails closed when port is 0' test called request() with an extra argument the helper ignores, then never asserted on res. Real coverage lives in the dedicated 'fail-closed before port resolution' describe block. |
||
Ajay Satish
|
9d700ec74f
|
feat(daemon): persist code agent startup (#255)
* feat(daemon): persist code agent startup * fix: complete all suggestions * fix: types for app config * chore: revert local origin * chore: format to single quotes * fix: duplicate headers * fix: isLocalSameOrigin rewriting issue --------- Co-authored-by: mrcfps <mrc@powerformer.com> |
||
|
monshunter
|
30f8036c9a
|
fix(web): make share-menu "Download as .zip" return the actual project tree (#341)
* fix(web): download project tree as zip from share menu The "Download as .zip" share action previously produced a single-file ZIP of the rendered HTML srcdoc. Add a daemon archive endpoint that bundles the on-disk project tree (scoped to the active top-level directory when applicable) and have FileViewer call it, falling back to the in-memory single-file ZIP on failure. UTF-8 filenames are preserved via RFC 5987 Content-Disposition. * fix(daemon,web): address PR #341 review — diagnostics, comments, more tests - Stat the archive root up-front so a missing/non-directory target surfaces a clear ENOENT/ENOTDIR ("does not exist" vs "is empty"), instead of being swallowed by the recursive walk and reported as empty. Distinguishes a deleted project from one with no archivable files for on-call diagnostics. - Document the DEFLATE level-6 choice in the archive builder so future maintainers don't have to guess at the speed/ratio trade-off. - Add a daemon test that the baseName preserves a multi-byte UTF-8 directory name (café-design), covering the path that feeds RFC 5987 filename* on the server. - Add a daemon test that a missing archive root surfaces ENOENT with a "does not exist" message, distinct from the empty-directory case. - Export archiveRootFromFilePath and archiveFilenameFrom for testing, and add web unit tests covering the Content-Disposition fallback chain (UTF-8 filename* → legacy quoted filename= → root slug → title slug, plus malformed-encoding fall-through). - Replace the CJK example string in code comments and tests with a Latin-extended example (café-design) so the codebase stays English while still exercising multi-byte UTF-8 handling. |
||
Sid
|
648374d839
|
fix(platform): wrap cmd.exe shim invocations to survive /s /c quote stripping (#339)
PR #258 standardized agent spawning through `createCommandInvocation`, which on Windows wraps `.cmd` / `.bat` paths in `cmd.exe /d /s /c <line>` and quotes each argument with cmd-style doubled quotes. PR #232's follow-up fix for `shell:true` was lost in that refactor, and the new shape has its own quoting bug on argv-style spawn: 1. cmd.exe `/s /c` strips exactly one leading and one trailing `"` from the rest of the command line. 2. Node, with `windowsVerbatimArguments` unset, escapes each argv element using CommandLineToArgvW rules — so the inner `"path with space"` ends up surfacing to cmd.exe with an extra layer of `\"` escaping that cmd doesn't understand. Together these collapse `"C:\Users\Ethical Byte\...\codex.CMD" --help` into `C:\Users\Ethical Byte\...\codex.CMD --help` with no quoting preserved, and cmd.exe parses the first space as a token boundary — "`Ethical` is not recognized as an internal or external command." See issue #315 for the full repro. The fix mirrors what Node's own `child_process.spawn({ shell: true })` does internally: wrap the entire joined command line in an extra `"…"` and set `windowsVerbatimArguments: true`. The outer wrap absorbs the `/s /c` strip, leaving inner per-arg quoting intact, and the verbatim flag tells Node to pass argv through to CreateProcess unchanged. Changes: - `packages/platform/src/index.ts` - Extend `CommandInvocation` with optional `windowsVerbatimArguments`. - Extract the cmd.exe shim builder into `buildCmdShimInvocation` and apply the outer wrap + verbatim flag in both `createCommandInvocation` and `createPackageManagerInvocation`. - Forward the flag through `spawnBackgroundProcess` and `spawnLoggedProcess`. - `apps/daemon/src/server.ts` — agent spawn forwards `invocation.windowsVerbatimArguments`. This is the call site that hit #315 in the wild (Codex CLI `.CMD` shim, user dir with space). - `tools/pack/src/win.ts` — `runPnpm` and `runNpmInstall` forward the flag through `execFileAsync`. Affects the Windows packaged-build pipeline when run from a path with spaces. - `tools/dev/src/index.ts` — `runLoggedCommand` accepts and forwards the flag; `buildDesktop` propagates it from `createPackageManagerInvocation`. Affects local dev on Windows. Tests: - 9 new unit tests in `packages/platform/src/index.test.ts` stub `process.platform` so both Windows and POSIX branches run on every CI runner. Coverage: - POSIX pass-through. - Windows non-shim binary pass-through. - `.CMD` shim with spaces in the binary path (the #315 repro). - `.bat` shim parity. - Argv elements with spaces alongside the shim path. - Argv elements without whitespace stay unquoted. - `process.env.ComSpec` fallback. - `npm_execpath` short-circuit (cross-platform). - POSIX pnpm pass-through. - Windows pnpm wrapped through cmd.exe. Closes #315. |
||
|
lefarcen
|
e399593528
|
fix(daemon): increase projectUpload limit from 20MB to 200MB (#319)
The /api/projects/:id/upload endpoint used the projectUpload multer handler with a 20MB limit. Large design files (e.g., 47MB PPTX) were rejected with HTTP 413 Payload Too Large. Bumping to 200MB rather than the originally-proposed 100MB so the cap comfortably covers the heaviest design assets we realistically see — high-fidelity PPTX exports, raw camera images, and PDF source files — without forcing chunked-upload work on the project-side path. The daemon listens on 127.0.0.1 only and writes via multer's disk storage backend, so per-request memory pressure stays bounded. importUpload (separate /api/import/claude-design path) keeps its 100MB limit; the two endpoints don't have to match and bumping import in lockstep is a separate decision. Closes #318 |
||
NickSeagull's Bot
|
71370fd9dd
|
feat(daemon): add OD_BIND_HOST env var and --host flag for interface binding (#328)
Adds a `host` option to `startServer()` (sourced from `OD_BIND_HOST` env or `--host` CLI flag, defaulting to `'0.0.0.0'` to preserve existing behavior) and wires it through `app.listen(port, host, ...)`. This lets operators restrict the daemon to a single network interface — e.g. a Tailscale IP — for tailnet-only deployments. Without this knob the daemon always bound to all interfaces with no override path. Backward-compatible: the default `'0.0.0.0'` matches current behavior. The reported URL switches from the hardcoded `127.0.0.1` to the actual bound address when a specific interface is requested. Co-authored-by: Paperclip <noreply@paperclip.ing> |
||
Nagendhra Madishetti
|
8897cb85be
|
feat(deploy): add /api/projects/:id/deploy/preflight for pre-upload inspection (#320)
* feat(deploy): add /api/projects/:id/deploy/preflight for pre-upload inspection
Today the deploy flow is a single POST that builds the file set, ships
it to Vercel, and waits for the public URL to come up. The user has no
visibility into what is actually being uploaded until the operation
either succeeds or fails with a generic 400. There is no way to tell
"my background image is missing" from "my deploy will exceed Vercel
quotas" without rolling the dice and waiting up to ~110 seconds.
Add a preflight endpoint that builds the plan, runs an analyzer, and
returns a typed report without touching Vercel. The endpoint is purely
additive, costs no network round-trips, and gives the UI everything
needed to render a confirm-before-deploy summary.
Implementation:
- packages/contracts/src/api/projects.ts: new public types
DeployPreflightRequest, DeployPreflightResponse, DeployPreflightFile,
DeployPreflightWarning, and the closed enum DeployPreflightWarningCode
with values broken-reference, invalid-reference, large-asset,
large-bundle, large-html, external-script, external-stylesheet,
no-doctype, no-viewport.
- apps/daemon/src/deploy.ts: refactor buildDeployFileSet so the
walk-and-collect logic lives in a non-throwing buildDeployFilePlan
that returns { entryPath, html, files, missing, invalid }.
buildDeployFileSet keeps its existing semantics by delegating to the
plan and throwing when missing or invalid is non-empty, so every
current caller and test is unchanged. Add analyzeDeployPlan that
walks the entry HTML once via parseHtmlTags + parseHtmlAttributes
(no new parser) and emits the warning vocabulary above. Add
prepareDeployPreflight that combines the plan and the analyzer and
returns the public DeployPreflightResponse shape. Soft thresholds
(4 MiB per asset, 75 MiB bundle, 1 MiB entry HTML) are exported as
DEPLOY_PREFLIGHT_LARGE_*_BYTES constants so tests and future tuning
use the same numbers.
- apps/daemon/src/server.ts: new route POST /api/projects/:id/deploy/
preflight. Same body shape as POST /deploy ({ fileName, providerId? }).
Returns DeployPreflightResponse on success. Validation errors map to
BAD_REQUEST, missing entry file maps to FILE_NOT_FOUND, exactly like
the existing deploy POST so the client error-handling stays uniform.
Tests (apps/daemon/tests/deploy.test.ts, 14 new cases under a new
"deploy plan and analyzer" describe block):
- buildDeployFilePlan returns files plus missing and invalid lists
without throwing.
- buildDeployFileSet still throws on missing/invalid, preserving the
pre-refactor contract.
- Analyzer unit tests for broken-reference, invalid-reference,
no-doctype, no-viewport, external-script, external-stylesheet,
protocol-relative external script, large-asset (per-file but not
entry HTML), large-html (entry HTML threshold), and a healthy-input
case that produces zero warnings.
- Preflight integration tests: full payload shape (provider, entry,
files, totals, warnings) on a healthy project, broken-reference path
on a missing asset, and a non-HTML entry rejection.
Net diff: +439 / -8 across 4 files. No new dependencies, no public API
removals. The existing POST /deploy flow is byte-for-byte unchanged
because buildDeployFileSet is a thin wrapper around the new plan.
* fix(deploy): preflight review fixes (P2 doctype anchoring, P2 large-html source path, P3 logging, P3 jsdoc)
Address review feedback on PR #320:
P2 (Codex bot): the no-doctype check used /<!doctype\s+html/i with no
anchor, so a `<!doctype html>` substring inside a <script> template
literal or HTML comment would mask a missing real declaration. Switch
to `new RegExp('^\uFEFF?\s*(?:<!--[\s\S]*?-->\s*)*<!doctype\s+html', 'i')`
which anchors to the document prolog and accepts the HTML5-permitted
optional BOM, comments, and whitespace before the doctype.
P2 (Codex bot): the large-html warning hardcoded `path: 'index.html'`,
but `entry` in the preflight payload is the source project file (e.g.
`pages/landing.html`) and other warnings already use source paths.
Switch the warning path to `entryPath` so file-targeted UI actions
deep-link back to a real project file.
P3 (lefarcen): add a JSDoc block alongside the inline TS annotation
on analyzeDeployPlan so the function follows the JSDoc-first style
of the rest of the file. The TS inline annotation stays because
deploy.ts uses `// @ts-nocheck`, which makes JSDoc `@param`/`@returns`
non-authoritative for callers; both forms together give docs and
correct caller-side types.
P3 (lefarcen): the preflight route catch block now logs non-DeployError
exceptions via console.error before sending the generic 400, so
unexpected failures show up in the daemon log without leaking
internals to the client.
Tests:
- analyzeDeployPlan flags missing doctype even when a fake `<!doctype>`
lives inside a <script> string literal.
- analyzeDeployPlan accepts a doctype that follows a leading HTML
comment and BOM (HTML5 prolog forms).
- analyzeDeployPlan reports large-html against the source entry path
(e.g. `pages/landing.html`), not the deploy-renamed `index.html`.
38 deploy tests pass, 264 across the daemon package, typecheck clean.
---------
Co-authored-by: Nagendhra <nagendhra405@gmail.com>
|
||
|
Tom Huang
|
6fa2077651
|
feat(web): add pet companion with Codex hatch-pet integration (#296)
* feat(web): add pet companion with Codex hatch-pet integration
Introduces a customizable floating pet companion (overlay + entry-view rail
+ composer menu + dedicated Settings → Pets section) that supports built-in
pets, user customization (glyph/image/spritesheet), and one-click adoption
of pets packaged by the upstream Codex `hatch-pet` skill via a new
`/api/codex-pets` daemon endpoint. Vendors the unmodified `hatch-pet`
skill under `skills/hatch-pet/` and adds i18n strings across all locales.
Co-authored-by: Cursor <cursoragent@cursor.com>
* feat(scripts): sync community Codex pets from public catalogs
Adds `pnpm sync:community-pets` which fetches all pets from
codex-pet-share.pages.dev (paginated Supabase Functions API) and
j20.nz/hatchery (single-shot JSON), then writes each one as
`<id>/pet.json` + `<id>/spritesheet.webp` under
`\${CODEX_HOME:-\$HOME/.codex}/pets/`. The existing daemon
`codex-pets` registry already scans that folder, so synced pets
appear under Settings → Pets → Recently hatched and adopt with one
click — no manual upload. Supports --source/--out/--force/--limit
flags and validates magic bytes so HTML error pages never end up
masquerading as `.webp` files.
Co-authored-by: Cursor <cursoragent@cursor.com>
* fix(daemon): tighten codex-pets validation and document vendoring
- sanitizeId now rejects ids that still contain `..` after collapsing,
closing a defensive gap on the path-traversal guard for the
`/api/codex-pets/:id/spritesheet` route.
- listCodexPets emits the sanitised folder name as the public id so the
download route resolves directly against the on-disk folder, even when
`manifest.id` differs (manual drops, sanitiser-touched manifests).
- Drop `@ts-nocheck` from `codex-pets.ts`; module is now strict-typed
with explicit interfaces, an unknown-narrowed JSON.parse path, and a
`pickString` helper guarding manifest fields one by one.
- Restrict the spritesheet response CORS header to sandboxed-iframe
callers (Origin: null) instead of unconditional `*`, matching the
existing raw-file route pattern. Same-origin web traffic does not
need the header (web proxies `/api/*` through the daemon).
- Add `skills/hatch-pet/README.md` explaining the vendoring trade-off,
provenance, and re-sync procedure.
- Add `docs/codex-pets.md` covering where pets live, how to populate the
registry without Codex installed, and the manifest contract.
Generated-By: looper 0.4.0 (runner=fixer, agent=claude-code)
* fix(i18n): add pet.* keys to Hungarian locale
Hungarian locale was added on main after this branch diverged, so the new
pet.* dictionary keys never landed there and tsc -b reports hu's Dict as
incomplete once main is merged in.
Generated-By: looper 0.4.0 (runner=fixer, agent=claude-code)
* feat(web): atlas-driven pet animations + bundled community pets
Builds on the existing pet companion (#296) with a richer animation
loop, a curated set of community pets that ship with the repo, and a
one-click sync into ~/.codex/pets/.
- Atlas-mode rendering: PetSpriteFace can now play the full Codex 8x9
sprite atlas and swap rows from a JS-driven frame index. PetOverlay
classifies pointer interactions (idle / hover / drag-direction /
long-idle waiting) and maps them to the matching atlas row, so the
pet waves on hover, runs on drag, and falls into a waiting pose
after 6s of stillness. Single-strip pets keep their existing CSS
steps() animation, with the steps timing fixed to jump-none so frame
cells line up on cell boundaries.
- Atlas adoption: PetSettings exposes both "Use full atlas (animated)"
and "Freeze to this row" — full mode keeps every row for the
interaction state machine, single-row mode crops one strip via the
existing canvas helper. New prepareCodexAtlas downscales the atlas
to a localStorage-friendly PNG while preserving the grid layout.
- Settings tabs: pet sources are now split into Built-in / Custom /
Community tabs so each origin gets its own dedicated surface.
- Bundled pets: scripts/bake-community-pets.ts seeds a curated set
(clippit, dario, nyako-shigure, slavik, trump, tux, yelling-dario,
yorha-sit-2b) into assets/community-pets/. The daemon scans this
alongside the user's ~/.codex/pets/ root, with user pets winning
when ids collide. CodexPetSummary gains a `bundled` flag so the UI
can tag those cards with a "Bundled" pill.
- One-click community sync: daemon-side port of sync-community-pets
exposed via POST /api/codex-pets/sync. Returns the same
wrote/skipped/failed/total summary the CLI prints. Web Pet settings
surface this as a "Download community pets" button under the
Community tab.
- Avatar dropdown + hide rail: EntryView's avatar button is now a
small menu (mirrors the project-view AvatarMenu) with toggles for
hiding/showing the pet rail and opening Settings. PetRail gets a
matching × button for the same hide flow.
- Locales: 7 new pet.* keys for tabs, sync, hide/show, atlas full
mode, and the Bundled pill — translated into all 13 supported
locales.
Typechecks pass across all workspace packages; daemon + web vitest
suites stay green.
Co-authored-by: Cursor <cursoragent@cursor.com>
* feat(web): bundled-pets built-in tab, ambient atlas animations, and community sync button
The Built-in tab now sources its catalog from the bundled spritesheets
at `assets/community-pets/` instead of the eight emoji placeholders that
felt boring next to the Codex hatch-pet atlases.
- Daemon: `listCodexPets` flags `bundled: true` by curated-set membership
in `assets/community-pets/`, not by which folder the sprite happened to
be read from. Previously a fully-synced user inbox preempted every
bundled id and left the tab empty.
- Settings → Pets → Built-in renders the same sprite-card grid as
Community, filtered by `bundled: true`, and reuses the existing
`adoptCodexPet` flow. Community tab filters to non-bundled so the
curated set never appears twice.
- Community tab gains the long-promised "Download community pets"
trigger that calls `/api/codex-pets/sync` and shows an inline status
line for the run summary. Strings already existed in every locale; we
just plumbed the button.
- `PetOverlay` gets ambient atlas-row choreography — while idle, the
overlay occasionally swaps `idle` for a random non-idle row (wave /
hop / look) so the pet doesn't feel frozen. User gestures cancel the
beat and take over instantly. `pickAmbientRow` lives next to
`pickAtlasRow` so both row pickers share the fallback discipline.
- One-shot `migrateCustomPetAtlas` heals configs adopted before the
overlay learned row switching by re-downloading the full spritesheet
so hover / drag / ambient variety light up on next launch.
- `BUILT_IN_PETS` is now an empty array (the type stays for backwards
compat); legacy configs whose `petId` still points at an emoji id
(`mochi`, `pixel`, …) fall back to the user's custom slot in
`resolveActivePet` so the overlay never renders blank.
- i18n: refresh `pet.tabBuiltInHint` (drop "emoji companions") and add
`pet.builtInEmpty` across all locales.
Co-authored-by: Cursor <cursoragent@cursor.com>
---------
Co-authored-by: Cursor <cursoragent@cursor.com>
|
||
|
Sid
|
5edd852cfa
|
fix(daemon): restore startServer Promise contract — return url / { url, server } (#268)
* fix(daemon): restore startServer Promise contract — return url / { url, server }
PR #258 ("Standardize agent communication via stdin and remove
Windows-specific shims logic") removed the Promise wrap around
`startServer`'s `app.listen()` call. Two regressions shipped to main as
a result:
1. **`returnServer: true` returns just `server`, not `{ url, server }`.**
Both `apps/daemon/sidecar/server.ts:65` and
`apps/daemon/tests/version-route.test.ts:10` cast the return value to
`{ url, server }` and read `started.url`. Post-#258 those reads land
on `undefined`, which becomes `"undefined/api/version"` URLs at
runtime — the sidecar can no longer report its bound port to the
parent process, and `version-route.test.ts` fails with
`ERR_INVALID_URL`.
2. **`returnServer: false` returns nothing.** `apps/daemon/src/cli.ts:77`
calls `startServer({ port }).then(url => console.log(\`[od] listening on \${url}\`))`
— post-#258 that prints `[od] listening on undefined`, and the
subsequent `spawn(opener, [url], …)` to open the browser also gets
`undefined`. The daemon still binds, but the two paths that *consume*
the URL are silently broken.
Both stem from the same change: the new code synchronously kicks off
`app.listen` and returns either the server object or `undefined` before
`listen`'s callback fires, so neither the resolved port nor the URL
string ever reach the caller.
This restores the pre-#258 Promise wrap, which already had the right
shape for all three call sites:
```ts
return await new Promise((resolve) => {
const server = app.listen(port, () => {
const url = `http://127.0.0.1:${server.address().port}`;
resolve(returnServer ? { url, server } : url);
});
});
```
Behavior contract restored:
| `returnServer` | resolves with |
|---|---|
| `false` (default) | `string` — `http://127.0.0.1:<port>` |
| `true` | `{ url: string, server: http.Server }` |
The Promise resolution waits for `listen`'s callback so `port=0`
(ephemeral) callers always see the actual bound port, never the
placeholder.
Test coverage: `tests/version-route.test.ts` already exercises the
`{ url, server }` path via `await fetch(\`\${baseUrl}/api/version\`)`.
After this fix the suite is back to 15 files / 225 tests green
(2 of those — both in `version-route` — failed against current main).
Verification:
- `pnpm --filter @open-design/daemon test` — 15 files / 225 tests pass
- `tsc -p apps/daemon/tsconfig.json --noEmit` — clean
- `tsc -p apps/daemon/tsconfig.tests.json --noEmit` — clean (was failing
on the `as { url, server }` cast before)
* fix(daemon): reject startServer Promise on bind failure / null address
Folds in the two non-blocking improvements @lefarcen flagged on the
review of the parent commit:
1. **Explicit `error` event listener.** `app.listen` throws synchronously
for some failure modes (e.g. invalid port) and emits an `error` event
for others (EADDRINUSE on certain Node versions, EACCES, EADDRNOTAVAIL).
Without an `error` listener the returned Promise would hang forever
on the event-emitting paths instead of surfacing the failure to the
caller.
2. **Null-`address()` guard.** `server.address()` typing allows
`string | AddressInfo | null`. For TCP listeners it's always
AddressInfo, but if a future code path ever returns null (or a Unix
socket path string), the previous fallback `port` would still be
`0` for ephemeral-port callers and quietly produce a
`http://127.0.0.1:0` URL that fetches would fail on. Reject the
Promise instead with an explicit diagnostic.
Both keep the success path identical — `version-route.test.ts` still
exercises the same `{ url, server }` contract on `port: 0` and passes
unchanged. 15/15 daemon test files / 225/225 tests pass.
Refs review feedback on #268 (P2 suggestions, non-blocking).
* fix(daemon): drop
|
||
Caprika
|
0c00f241e7
|
Add preview comment attachments (#284) | ||
Rin
|
06b1492744
|
Standardize agent communication via stdin and remove Windows-specific shims logic (#258)
This change moves Claude Code and GitHub Copilot CLI to stdin-based prompt delivery, aligning them with other agents. By doing so, we can bypass OS command-line length limits on Windows and set shell: false universally, improving security and robustness. Integration of createCommandInvocation ensures that Windows npm shims (.cmd/.bat) are correctly handled while keeping shell: false, addressing the P1 warning from the automated review and maintaining consistent execution across platforms. |
||
|
lefarcen
|
c2b3d737f2
|
fix: make max_tokens configurable (closes #29) (#78)
* fix(web,daemon): make max_tokens configurable (closes #29) BYOK users on custom Anthropic-compatible providers (e.g. Xiaomi MiMo) hit the hardcoded 8192 cap and saw artifacts truncated mid-stream. - AppConfig.maxTokens with Settings input (EN/CN + 8 other locales) - ProxyStreamRequest.maxTokens contract field - anthropic, anthropic-compatible, and openai-compatible providers all forward cfg.maxTokens - /api/proxy/anthropic/stream and /api/proxy/stream payloads honor it, defaulting to 8192 when unset so prior clients are unaffected Original sketch by @mashu in #78 (50a9d14); rebased to the apps/web layout and extended to the proxy paths actually used when baseUrl is set, which is where #29's user actually traffics. * feat(web): per-model max_tokens defaults Adds a hand-maintained MODEL_MAX_TOKENS table (Claude 4.5 line → 64k, mimo-v2.5-pro → 32k) and an effectiveMaxTokens helper layered over the override field added in |
||
|
Tom Huang
|
1edab990bb
|
feat(craft): add brand-agnostic craft references + Refero-derived lint rules (#225)
* feat(craft): add brand-agnostic craft references and refero-derived lint rules Introduce `craft/` as a third top-level content axis alongside `skills/` and `design-systems/`, holding universal (brand-agnostic) craft rules that apply on top of any DESIGN.md. Skills opt in via a new `od.craft.requires` front-matter array; the daemon resolves the slug list and injects the matching files between DESIGN.md and the skill body in the system prompt. Initial vendor (MIT, adapted from referodesign/refero_skill): typography craft, color craft, anti-ai-slop. Pilot wired on saas-landing. Extend the existing lint-artifact pass with two refero-derived rules: - P0 ai-default-indigo — solid #6366f1 / #4f46e5 / #4338ca / #8b5cf6 as accent (not just gradients) is the most-reported AI tell. - P1 all-caps-no-tracking — `text-transform: uppercase` rules without ≥0.06em letter-spacing. The craft loader silently drops missing files so a skill can forward-reference future sections (e.g. `motion`) without breaking. * fix(daemon): skip :root token blocks in ai-default-indigo lint The ai-default-indigo P0 check scanned the whole HTML for the raw hex, so brands that intentionally encode indigo as `--accent: #6366f1` in :root and consume it via var(--accent) downstream were flagged as AI-default — a false positive that forced the agent to "fix" valid output. Strip :root token-definition blocks (including attribute-selector theme variants) before scanning, mirroring the existing pattern used by the raw-hex P1 check. Hex still flagged when it appears in component rules or inline styles. * docs(craft): address PR #225 P3 review feedback - craft/README.md: explain why missing craft sections are silently dropped (forward-compatibility) instead of surfacing a warning. - craft/typography.md: ground the 0.06em ALL CAPS tracking floor in Bringhurst-derived typographic practice rather than presenting the threshold as unattributed. - craft/color.md: cover the edge case where a brand's DESIGN.md intentionally encodes indigo as --accent — `var(--accent)` uses remain unflagged because the linter only inspects hardcoded hex. - docs/skills-protocol.md: link the "missing files dropped silently" note back to craft/README.md for the canonical slug list and the rationale behind the choice. * fix(craft): address PR #225 P0 review feedback - tools/pack: copy `craft/` into the packaged resource root alongside `skills`, `design-systems`, and `frames`, so the `od.craft.requires` integration isn't a silent no-op when the daemon resolves `${OD_RESOURCE_ROOT}/craft` in packaged builds. - packages/contracts: add `craftRequires?: string[]` to `SkillSummary` (and therefore `SkillDetail`) so the field that `listSkills()` already returns and `/api/skills(/:id)` already serializes via `...rest` is part of the documented web/daemon contract instead of leaking through as an untyped property. - apps/daemon/lint-artifact: expand the indigo token-strip pass to cover selector lists containing `:root` (e.g. `:root, [data-theme="light"]`) and any rule whose body is custom-property-only (e.g. a `[data-theme="dark"] { --accent: ... }` theme variant). Real component rules with a hardcoded indigo are still preserved so the P0 finding still fires; tests cover the new selector-list and theme-variant cases. * fix(craft): address PR #225 follow-up review feedback - lint-artifact: scope the indigo token-strip to <style> blocks so the rule-shaped regex no longer captures leading `<style>` text into the selector (which broke `:root` recognition for token blocks that mix `color-scheme`/etc. with `--accent`). Run the strip on the extracted CSS instead, with a regression covering `:root { color-scheme: light; --accent: #6366f1 }`. - lint-artifact: tighten the custom-property-only exemption to global theme-scope selectors (`:root`, `html`, `body`, bare attribute selectors like `[data-theme="dark"]`). Component-local rules such as `.cta { --cta-bg: #6366f1 }` are no longer exempted, so an agent cannot launder default indigo through a local var. Regression test added. - craft/anti-ai-slop.md: stop claiming every rule below is enforced by the linter; only several are. The unenforced rules (standard Hero→Features→Pricing→FAQ→CTA flow, decorative blob/wave SVG backgrounds, perfect symmetry) are now flagged inline as "(guidance, not auto-checked)" so the contract with the lint surface stays honest. * fix(daemon): tighten lint-artifact iteration and :root token gating - all-caps-no-tracking: iterate every <style> block. The previous check called `exec` once on a non-global regex, so an artifact whose offending uppercase rule sat in a second <style> block (e.g. a reset block followed by a components block) slipped past. Switch to `matchAll` and break across both loops once a violation is found. Regression test covers a second-block uppercase rule. - ai-default-indigo: stop unconditionally exempting any selector list containing `:root`. The exemption now requires both conditions to hold: every selector in the list is global theme scope AND the body is token-shaped (CSS custom properties or the `color-scheme` keyword). So `:root { background: #6366f1 }` and `:root, .cta { --cta-bg: #6366f1 }` no longer launder a hardcoded indigo through the strip pass. Regression tests cover both bypass shapes. * fix(daemon): scope theme-attr exemption and strip CSS comments in token blocks Address PR #225 review feedback on `ai-default-indigo`: - The bare-attribute branch of `selectorListIsGlobalThemeScope` accepted any `[attr=...]` selector, so a custom-property-only rule on a component/state attribute (e.g. `[data-variant="primary"]`, `[aria-current="page"]`) was treated as a global theme block and stripped before the indigo scan — exactly the component-local indigo laundering this lint is meant to catch. Restrict the exemption to a small allowlist of known theme switches: `data-theme`, `data-color-scheme`, `data-mode`. - `stripTokenBlocksFromCss` split rule bodies on `;` and matched each fragment from the start, so a token block whose body contained a normal CSS comment such as `:root { /* brand accent */ --accent: #6366f1; }` produced a fragment beginning with the comment, failed `isTokenShapedDeclaration`, and the rule was left in scope of the indigo scan — a false P0 on a legitimate token definition. Strip CSS comments before splitting/classifying declarations. Add regression coverage: arbitrary component/state attribute selectors still trip `ai-default-indigo`; `data-color-scheme` theme variants stay exempted; `:root` token blocks with leading, trailing, and between-declaration CSS comments are recognized. * fix(daemon): strip CSS comments and recognize tokens nested in at-rules The all-caps-no-tracking scan ran against raw `<style>` content, so a commented-out rule like `/* .eyebrow { text-transform: uppercase; } */` matched `upperRe` and emitted a P1 for CSS the browser ignores. Strip CSS comments from the style body before structural matching. `stripTokenBlocksFromCss` only matched flat `selector { body }` rules, so a media-query-wrapped token block like `@media (prefers-color-scheme: dark) { :root { --accent: #6366f1 } }` had its outer `@media` rule treated as the selector/body pair and the inner `:root` token block was never stripped, producing a P0 false positive on legitimate responsive theme CSS. Tighten the body alternation to `[^{}]*` so the regex matches innermost rules and recognizes the inner `:root` block directly while preserving the outer at-rule wrapper. * fix(daemon): align ai-default-indigo list with documented cardinal sins The lint's AI_DEFAULT_INDIGO subset omitted #3730a3 and #a855f7, which craft/anti-ai-slop.md lists as P0-blocked solid accents. An artifact could hard-code one of those documented colors as a button fill and slip past the indigo scan unless it happened to be inside a gradient. Bring the lint set to the exact list documented in the craft doc, and tighten the doc's wording from "etc." to an explicit enumeration that points at AI_DEFAULT_INDIGO so the prompt contract and daemon behavior stay in sync. Add regression tests pinning each newly-included hex. * fix(daemon): tighten theme-scope selector and scan inline ALL CAPS The theme-scope exemption used to accept any attribute on `:root`, `html`, or `body` (e.g. `:root[data-variant="primary"]`), letting an agent launder default indigo through a component/state attribute and slip past the `ai-default-indigo` lint. The prefixed branches now require the attribute name to be one of GLOBAL_THEME_ATTRIBUTES, matching the bare-attribute branch. The `all-caps-no-tracking` rule only iterated `<style>` blocks, so inline declarations like `<span style="text-transform: uppercase">` produced no finding even though craft/typography.md treats the ≥0.06em tracking floor as having no exceptions. Added a second scan over `style="..."` attributes that runs the same letter-spacing check and dedupes against the existing `<style>`-block finding so the agent gets a single corrective signal per artifact. * fix(daemon): align uppercase tracking px floor with the 0.06em rule The previous absolute fallback (>=1.5px) was stricter than the craft rule it enforces. `font-size: 12px; letter-spacing: 1px` is 0.083em — above the 0.06em floor — but 1.5px would reject it and trigger an unnecessary correction loop on compliant small-label CSS. Extract `hasAdequateUppercaseTracking`: read `font-size` from the same rule body and compare px tracking against `fontSize * 0.06`; fall back to a conservative >=1px floor when font-size is inherited (covers the default 16px body where 1px ≈ 0.0625em). Apply the helper to both the <style>-block scan and the inline-style scan, and add 12–14px label tests in both branches. * fix(daemon): treat rem letter-spacing as absolute, not per-element em `rem` was previously folded into the same branch as `em` and accepted at the 0.06 threshold. But `rem` is relative to the root font-size (16px default), not the element's own font-size, so on a 48px heading `letter-spacing: 0.06rem` resolves to 0.96px — about 0.02em of the element, well below the 0.06em rule the lint enforces. Convert rem to absolute px through the 16px root assumption and reuse the same px-vs-element-font-size resolution: same-rule `font-size: <n>px` gives an exact `n * 0.06` floor; otherwise the conservative >=1px fallback applies. Add regression tests for 48px headings with 0.06rem tracking (must flag) plus the 16px-element and rem-floor matches that must keep passing, in both <style>-block and inline-style branches. * fix(daemon): resolve var() refs in uppercase tracking lint `hasAdequateUppercaseTracking` only matched literal numeric values, so a tokenized rule like `letter-spacing: var(--caps-tracking)` — exactly the pattern the craft prompt steers artifacts toward — was falsely reported as `all-caps-no-tracking`. Extract `--name: value` declarations from global theme scopes (`:root`, `html`, theme-attribute selectors) once per artifact, then expand simple `var(--name)` (and `var(--name, fallback)`) references in the inspected rule body before applying the existing 0.06em / px-floor / rem-conversion logic. References without a matching token and no fallback stay in place, preserving the conservative "missing tracking" finding. * fix(daemon): resolve rem and var() font-size in uppercase tracking lint Previously the px-vs-element-font-size resolution only matched `font-size: <n>px`. Any rem-based or tokenized display size fell through to the lenient `>= 1px` body-text fallback, so an artifact emitting `.display { font-size: 3rem; text-transform: uppercase; letter-spacing: 1px; }` (a ~48px heading with a 2.88px floor) slipped past the lint that this helper exists to enforce. Resolve `rem` font-size via the same root-font assumption already used for tracking, and treat any explicitly declared but unresolvable unit (`em`, `%`, `calc(...)`, an unresolved `var(...)`) conservatively — refuse the lenient fallback so the rule must use either an `em` letter-spacing or a verifiable px/rem font-size. `var()` font-size declarations resolve through the existing `resolveCssVars` pass before the size scan runs, so the same fix catches the tokenized-display-size pattern (`--display-size: 3rem`). * fix(daemon): parse declarations to ignore custom-prop names in uppercase tracking lint The hasAdequateUppercaseTracking and resolveFontSizePx helpers used substring regexes against the rule body, so a token-name declaration such as `--letter-spacing: 0.08em` or `--display-font-size: 48px` could satisfy the `letter-spacing` / `font-size` checks even though it has no rendered effect — letting actual ALL-CAPS-without-tracking rules slip past the P1 lint. Parse the declaration list, compare exact property names, and skip declarations whose property starts with `--`. Adds regression tests covering token-name letter-spacing (style-block + inline) and a token-name font-size masking the bail-out branch. * fix(daemon): scope indigo token exemption to --accent only Previously stripTokenBlocksFromCss removed every custom-property-only global theme block before the ai-default-indigo scan, which let a laundered indigo token like `:root { --primary: #6366f1 }` consumed via `var(--primary)` slip past the lint. The craft contract is that the only escape hatch is encoding indigo as the design system's `--accent` token; any other token name is still the LLM-default color hidden behind an arbitrary name. Narrow the strip pass so a non-`--accent` token whose value carries an AI-default indigo hex keeps the rule in scope, and add regression tests for `--primary` / `--button-bg` global tokens feeding a CTA, including the at-rule and theme-attribute variants. * fix(daemon): model CSS cascade in tracking lint and detect blue→cyan trust gradients Address PR #225 review feedback (3 comments): - `letter-spacing` / `font-size` selection now picks the LAST matching declaration in the rule body, modeling CSS source-order cascade. `.eyebrow { letter-spacing: 0.08em; letter-spacing: 0.02em }` renders the noncompliant 0.02em the browser actually shows; the previous first-match behaviour silently passed it. - `extractCssTokens` now records every distinct value seen for a token across global theme scopes, and `hasAdequateUppercaseTracking` enumerates each combination so a default-theme value below the floor cannot be rescued by a scoped override that happened to be parsed later (`:root { --caps-tracking: 0.02em }` + `[data-theme="dark"] { --caps-tracking: 0.08em }` now fires). - New `trust-gradient` P0 rule pairs blue/sky tokens against cyan tokens in `linear-gradient(...)` bodies so `blue→cyan` two-stop trust gradients (documented as a cardinal sin in `craft/anti-ai-slop.md`) are actually enforced — both the hex form (`linear-gradient(90deg, #3b82f6, #06b6d4)`) and the keyword form (`linear-gradient(90deg, blue, cyan)`). Adds 11 regression tests covering each path (cascade override in <style> and inline form, font-size cascade shifting the floor, both orderings of the conflicting-token cascade, the don't-over-fire case when every theme value clears the floor, hex / keyword / sky variants of the trust gradient, and the don't-double-fire case when purple-gradient already caught a mixed gradient). * fix(daemon): apply per-scope cascade in extractCssTokens When the same CSS custom property is declared more than once inside a single rule body (e.g. `:root { --caps-tracking: 0.02em; --caps-tracking: 0.08em }`), CSS source-order cascade collapses to the last value; the earlier declaration never reaches any element. `extractCssTokens` was treating intra-scope duplicates as simultaneous theme alternatives, so `hasAdequateUppercaseTracking` enumerated the stale 0.02em and emitted a spurious all-caps-no-tracking finding. Collapse duplicate token declarations within a rule body to the last value before merging into the cross-scope distinct-value map. Cross-scope overrides (separate `:root` and `[data-theme]` rules) remain preserved as distinct values so the conservative theme-cascade check still fires when ANY applicable theme renders below the floor. * fix(daemon): scope tracking lint to innermost rules and per-theme tokens Restrict the upperRe body alternation to [^{}]* so the regex matches innermost CSS rules and skips at-rule wrappers — an outer @media or @supports could otherwise capture as a single rule whose selector was the at-rule and whose body began with the inner selector token, masking the same-rule font-size and letting noncompliant tracking on large headings slip through the lenient inherited-size fallback. Replace the by-name-distinct-values token map with per-scope token records and a buildResolvedThemes pass that materializes one effective map per theme. Paired token declarations now stay paired during evaluation, so theme variants like :root + [data-theme=dark] no longer generate cross-theme cartesian pairings (e.g. default-size + dark-track) that emit false positives on legitimate light/dark themes. --------- Co-authored-by: looper <looper@open-claude.dev> |
||
yyh6666
|
85714d58c8
|
fix: daemon OD_DAEMON_URL uses port 0 instead of actual allocated port (#240)
When tools-dev launches the daemon with OD_PORT=0 (letting the OS pick a free port), app.listen correctly resolves the actual port in its callback, but the enclosing closure still references the original `port` parameter (value 0). This causes OD_DAEMON_URL to be set to http://127.0.0.1:0 in the environment passed to agent subprocesses, breaking media generation and any daemon callback that relies on isLocalSameOrigin. Introduce a mutable `resolvedPort` variable, assigned from `actualPort` inside the listen callback. All downstream consumers (OD_DAEMON_URL, isLocalSameOrigin calls) now read the resolved value. Co-authored-by: yyh <yyh@test.cn> |
||
VitaminBFFM
|
a263a9c11e
|
fix(daemon): quote agent bin path when spawning with shell:true on Windows (#232)
When `useShell` is true (Windows + .cmd/.bat shim), Node.js spawns the agent via `cmd.exe /d /s /c "<bin> <escaped-args>"`. Node escapes argv items but does NOT quote the bin path itself — that is the caller's responsibility. If `resolvedBin` contains spaces (the common case for npm shims under user directories with a space in the account name, e.g. `C:\Users\First Last\AppData\...\claude.CMD`), cmd.exe parses up to the first space as the command and the rest as arguments, then fails with `The system cannot find the file 'C:\Users\First'` (or the localized equivalent). The agent process exits in ~0.1s with code 1 and the chat surfaces the cmd.exe error verbatim. The Node DEP0190 deprecation warning that fires alongside the original spawn is the same root cause documented from the runtime side. Wrap `resolvedBin` in double quotes when `useShell` is true so the full path stays a single token. The `/s` flag that Node already passes to cmd.exe handles the resulting `cmd /d /s /c ""path with spaces" args"` correctly. |
||
Aresdgi
|
59e4966dda
|
feat(version): add app version awareness (#204)
* feat(version): add app version awareness * fix(version): detect packaged sidecars across platforms |
||
PerishFire
|
f604ff1ec2
|
Add Windows beta packaging and release assets (#191) | ||
JiangDing
|
74344a370f
|
fix(daemon): add CORS header to raw project file endpoint (#140)
* fix(daemon): add CORS header to raw project file endpoint srcdoc iframes have a null origin. When preview HTML fetches component files from /api/projects/:id/raw/*, the browser blocks the response because the route returned no Access-Control-Allow-Origin header. Only respond with the header when the request Origin is the string "null" — the signature of a srcdoc iframe. Real cross-origin requests from other websites remain blocked, keeping the endpoint safe even if the daemon is deployed on a remote server. Fixes #134 * fix(daemon): add CORS header to raw project file endpoint srcdoc iframes have a null origin. When preview HTML fetches component files from /api/projects/:id/raw/*, the browser blocks the response because the route returned no Access-Control-Allow-Origin header. Only respond with the header when the request Origin is the string "null" — the signature of a srcdoc iframe. Real cross-origin requests from other websites remain blocked, keeping the endpoint safe even if the daemon is deployed on a remote server. Also add an OPTIONS preflight handler to future-proof the route for artifacts that may send custom request headers, and add test coverage for the null-origin allow / real-origin block behaviour. Fixes #134 --------- Co-authored-by: jiangding001 <jiangding001@ke.com> |
||
ccfuncy
|
6e9f3cda73
|
fix(providers): proxy Anthropic-compatible streams (#180) | ||
|
Tom Huang
|
3f266103b0
|
feat(media): port generation workflow onto main (#12)
Co-authored-by: Elian <elian@EliandeMacBook-Pro.local> |
||
|
Caprika
|
454e8373fb
|
[feat] Add Vercel self deploy flow (#167)
* Add Vercel self deploy flow * Fix Vercel deploy file state and nested assets * Add deploy hook script injection |
||
|
PerishFire
|
a40d817d28
|
Add mac packaged runtime and beta release flow (#170)
* feat(pack): add mac packaged runtime control plane * feat(pack): harden mac packaged runtime lifecycle Keep packaged state namespace-scoped, make daemon paths explicit through sidecar launch env, and add conservative desktop identity/logging fallbacks for local mac package validation. * feat(pack): add mac beta release flow * fix(pack): generate mac update feed fallback * fix(pack): write portable beta checksums * fix(pack): make beta artifacts portable * fix(pack): clean up mac install visuals * fix(pack): address packaged runtime review feedback |
||
nettee
|
3fb849d047
|
Fix chat runs surviving web disconnects (#146)
* fix chat runs surviving web disconnects * fix chat run create abort propagation Generated-By: looper 0.0.0-dev (runner=fixer, agent=openai/gpt-5.5) * fix daemon keepalive reconnect budget Generated-By: looper 0.0.0-dev (runner=fixer, agent=gpt-5.5) * fix daemon stream disconnect cancellation Generated-By: looper 0.0.0-dev (runner=fixer, agent=openai/gpt-5.5) * fix daemon stream abort cancellation race Generated-By: looper 0.0.0-dev (runner=fixer, agent=openai/gpt-5.5) * fix daemon run cancellation semantics * fix load * doc * 2 * add run refresh recovery * fix active run refresh status * fix reattach abort handling * fix * fix chat initial scroll * fix daemon start failures Generated-By: looper 0.2.7 (runner=fixer, agent=openai/gpt-5.5) * fix background run recovery Generated-By: looper 0.2.7 (runner=fixer, agent=openai/gpt-5.5) * fix stop run status Generated-By: looper 0.2.7 (runner=fixer, agent=openai/gpt-5.5) * fix background run recovery Generated-By: looper 0.2.7 (runner=fixer, agent=openai/gpt-5.5) * extract daemon run service * move prompt composition to daemon * fix prompt module resolution * fix project id generation * add project run status * add designs kanban view with awaiting_input status - add grid/kanban view toggle on Designs tab; persist choice in localStorage - introduce awaiting_input project display status (daemon-derived from unanswered <question-form>) so projects asking the user aren't shown as Completed; ordered between Running and Completed with amber accent - hide transient queued state from users: coerce queued/starting to running in daemon /api/projects projection and drop the queued kanban column - a11y polish on Designs cards: Space activation, aria-labels on delete, focus-visible outlines, reveal delete on focus-within and touch, prefers-reduced-motion handling - kanban layout uses flex sizing instead of viewport math; scoped icon- only pill button rule fixes view-toggle icon alignment --------- Co-authored-by: mrcfps <mrc@powerformer.com> |
||
|
Xinmin Zeng
|
132adac3bb
|
fix(daemon): preserve non-ASCII filenames on multipart upload (#166)
* fix(daemon): preserve non-ASCII filenames on multipart upload
multer@1 hands callers latin1-decoded multipart filenames, and
sanitizeName() then collapses every non-ASCII character to '_'. A
Chinese DOCX uploaded as 测试文档.docx therefore landed on disk as
____.docx, while the response shipped a latin1-mangled originalName
back to the client. The chat composer compared that to the UTF-8
File.name from the picker, missed the match, and reported "some
files could not be stored" even though the bytes were already on
disk.
Add decodeMultipartFilename() that round-trip-checks latin1->utf8
so genuine latin1 names aren't corrupted, and switch sanitizeName()
to keep \p{L}\p{N} (Unicode letters/digits) while still stripping
path separators, control characters, and Windows-reserved
punctuation. Apply the decoder to all three multer storage filename
callbacks so /api/upload, /api/import/claude-design, and
/api/projects/:id/upload all return UTF-8 originalNames.
Closes #144
* fix(daemon): address review feedback on filename decoder
Three changes from review:
1. decodeMultipartFilename now bails out early when any code point in
the input exceeds 0xFF. multer can hand us an already-decoded UTF-8
string when the client uses the RFC 5987 `filename*` parameter, and
the previous round-trip-only check would corrupt those names — for
example, a properly decoded `测试文档.docx` could be re-mangled into
`KՇc.docx` because the latin1-encoded bytes happened to form a
valid UTF-8 sequence. Code points above 0xFF can never appear in a
genuine latin1-decoding-of-utf8 input, so they're an unambiguous
signal that no further decoding should run.
2. decodeMultipartFilename now handles null/undefined defensively (it
was relying on multer always populating originalname; this just
makes the helper safer to reuse from other call sites).
3. The inline sanitiser regex in the `upload` and `importUpload`
multer instances is replaced with a direct `sanitizeName()` call,
matching what `projectUpload` already did. This keeps a single
source of truth for the allowed character set so future tweaks
only have to land in one place.
Adds two more tests to the existing sanitize-name suite covering the
above-0xFF early return and the null/undefined inputs (13/13 in that
file, 58/58 across the daemon).
|
||
pan
|
7294929490
|
Fix daemon project-root resolution when launched from src via tsx (#162)
resolveProjectRoot only stripped a 'dist' suffix from the caller's module directory, so when the daemon sidecar is started by tools-dev (which runs apps/daemon/src/server.ts directly through tsx) the computed PROJECT_ROOT pointed at apps/ instead of the repo root. That made SKILLS_DIR, DESIGN_SYSTEMS_DIR, STATIC_DIR and the default RUNTIME_DATA_DIR all resolve to non-existent paths, so /api/skills returned an empty list and the web "examples" page reported "No skills available. Is the daemon running?". Treat 'src' the same as 'dist' so resolution works for both the tsx source entry and the compiled build, and add a regression test for the src case alongside the existing dist/daemon-root cases. |
||
Tom
|
8f34e39b7b
|
feat(daemon): add pi coding agent adapter (#117)
* feat(daemon): add pi coding agent adapter Add pi (https://pi.dev) as a supported coding agent, using its --mode rpc JSON-RPC protocol over stdio for structured event streaming. Changes: - apps/daemon/pi-rpc.js: new RPC session handler that drives pi's --mode rpc protocol, translating typed agent events (text_delta, thinking_delta, tool_use, tool_result, usage, status) into the daemon's UI event format. Auto-resolves extension UI requests (fire-and-forget consumed, dialogs auto-approved) so pi stays unblocked in the headless web UI. Kills the process after agent_end since pi's RPC process is designed for multi-prompt sessions. - apps/daemon/agents.js: add pi agent definition with custom fetchModels (pi --list-models outputs to stderr, not stdout), 575+ models from 20+ providers, reasoning/thinking level support via --thinking flag, and streamFormat 'pi-rpc'. - apps/daemon/server.js: wire pi-rpc stream format to attachPiRpcSession; skip stdin.end() for pi-rpc since the RPC session manages stdin bidirectionally. - apps/daemon/acp.js: export createJsonLineStream for reuse by pi-rpc.js. - apps/daemon/pi-rpc.test.mjs: 19 unit tests covering model list parsing (TSV, dedup, edge cases), RPC event translation (text, thinking, tools, usage, compaction, retry), sendCommand wire format, extension UI auto-resolution. - e2e/tests/structured-streams.test.ts: add pi RPC tool_use/tool_result event mapping test alongside existing Claude/Copilot fixtures. Verified end-to-end: daemon /api/chat → pi RPC → SSE stream with status, text_delta, usage, and tool events. Live E2E test passes (OD_E2E_RUNTIMES=pi). All 59 project tests green. * refactor(daemon): migrate pi-rpc to TypeScript Follow upstream #118 TypeScript migration convention: rename pi-rpc.js → pi-rpc.ts and pi-rpc.test.mjs → pi-rpc.test.ts with @ts-nocheck header (same as all other daemon modules). Import paths remain ./pi-rpc.js per NodeNext module resolution. * fix(daemon): avoid duplicate usage events in pi-rpc handler Pi emits both message_end and turn_end per turn, both carrying usage data. Emitting from both handlers caused double-counting in the UI and any consumer that aggregates usage. Remove usage emission from the message_end branch since turn_end is the canonical per-turn usage source. Keep tool call extraction in message_end (unique data not available in turn_end). Add regression test confirming exactly one usage event is emitted when both message_end and turn_end carry usage for the same turn. Addresses Copilot P2 review on PR #117. * fix(daemon): scope pi RPC id counter per session, bump graceful shutdown Move nextRpcId and sendCommand inside attachPiRpcSession as local state, matching the pattern in acp.ts where nextId is scoped per session. Prevents RPC id collisions across concurrent /api/chat requests. Bump post-agent_end SIGTERM grace period from 2s to 5s and make it configurable via PI_GRACEFUL_SHUTDOWN_MS env var for resource- constrained machines. Add test confirming concurrent sessions get independent id sequences. * fix(daemon): wrap parser.feed in try-catch in pi-rpc Catch errors from parser.feed and route them through the existing fail() handler instead of letting them propagate as unhandled exceptions. |
||
|
Caprika
|
bd2e71c708
|
fix codex plugin disable env (#133) | ||
|
PerishFire
|
c6d11018a0
|
Refresh desktop integration control plane (#123)
* feat(dev): add desktop tools-dev control plane * refactor(sidecar): split Open Design contracts Move Open Design-specific sidecar protocol definitions into @open-design/contracts so sidecar and platform can remain descriptor-driven primitives. * refactor(daemon): organize package sources Keep daemon app code, tests, and sidecar entrypoints in separate package directories so each layer can be built and verified independently. * chore(repo): streamline maintenance entrypoints Centralize agent guidance by directory and reduce root command chains while preserving the existing build scope. * docs: translate agent guidance to English * fix(sidecar): tolerate stale IPC sockets Remove stale Unix socket files only after confirming no listener is active, so tools-dev can restart after unclean shutdowns. |
Renamed from apps/daemon/server.ts (Browse further)