Instead of injecting the full SKILL.md body as additionalContext, inject
"You must run the Skill(<name>) tool." — a more conventional way of
telling the agent to use the Skill tool for context loading.
Hash invalid session IDs before constructing dedup temp paths so crafted stdin values cannot smuggle traversal segments into recursive claim cleanup. Shared temp-path resolution now verifies the resolved target stays under tmpdir, and seen-skills tests cover both stable safe IDs and hashed invalid IDs.
Verified: bun test tests/session-start-seen-skills.test.ts
How to test: bun test tests/session-start-seen-skills.test.ts
Swarm-Agent: codex-path-traversal
Parallel PreToolUse hooks were racing on the same seen-skills file, causing 2-6x
duplicate injections. UserPromptSubmit hook didn't share dedup state at all.
Fix by implementing atomic per-skill claim files (Oracle session: fix-dedup-race-condition-plan):
1. In hooks/src/hook-env.mts, add tryClaimSessionKey(sessionId, kind, key) using
openSync(path, 'wx') for atomic exclusive create. Add listSessionKeys() to read
claim dir, syncSessionFileFromClaims() to derive the txt snapshot, and
removeSessionClaimDir() for cleanup. Claim files live in
<tmpdir>/vercel-plugin-<sessionId>-seen-skills.d/<skill-name>.
2. In hooks/src/patterns.mts, add mergeSeenSkillStates(...values) to union env var,
session file, and claim dir state. Add serializeSeenSkills(set) helper.
3. In hooks/src/pretooluse-skill-inject.mts, load merged state from env+file+claims.
Before injecting each skill, call tryClaimSessionKey — first process wins, others
get EEXIST and skip. Use same claim path for once-only keys (agent-browser warning,
vercel-env-help). Pass sessionId through InjectOptions.
4. In hooks/src/user-prompt-submit-skill-inject.mts, preserve sessionId from parsed
input, load merged seen state, pass sessionId to injection flow so it uses the
shared claim backend. Skills injected via promptSignals now visible to PreToolUse.
5. In hooks/src/session-end-cleanup.mts, add rmSync deletion of claim directories.
6. Update CLAUDE.md dedup contract to document atomic claim mechanism.
7. Bump .plugin/plugin.json to 0.6.0, rebuild.
- Replace tsc+cp+rm pipeline with tsup (7ms builds, no intermediate dir)
- tsup compiles hooks/src/*.mts directly to hooks/*.mjs
- tsconfig.json retained for type-checking only (noEmit)
- Remove all .sh references from tests, docs, and hooks.json
- Fix glob pattern test to handle tsup's output format
Hook source now lives in hooks/src/*.mts and compiles to hooks/*.mjs via tsc.
The formatOutput() return type is checked against SyncHookJSONOutput from
@anthropic-ai/claude-agent-sdk, preventing unknown fields in hookSpecificOutput
that Claude Code's strict Zod validation would silently reject.
Root cause: skillInjection metadata was a sibling key in hookSpecificOutput,
which Claude Code rejected as "Hook JSON output validation failed: Invalid input".
All skill injections were silently discarded. Now skillInjection metadata is
embedded as an HTML comment inside additionalContext.
- Add hooks/src/*.mts (TypeScript source for all 6 hook modules)
- Add hooks/tsconfig.json (strict, NodeNext, compiles .mts → .mjs)
- Add build:hooks script and pre-commit hook for auto-compilation
- Install @anthropic-ai/claude-agent-sdk as devDependency for types
- Bump plugin version to 0.4.0
- Add greenfield project detection to session-start profiler
- Add profiler boost (+5 priority) for likely skills in PreToolUse
- Remove zzz-test-redos-pattern test skill
- Update golden fixtures and snapshots
- Add agent-browser to ecosystem graph
- Add new golden fixture tests
Create hooks/vercel-config.mjs that reads vercel.json and maps top-level keys (rewrites, redirects, headers, crons, functions) to skill slugs, so skills are injected when the project uses those Vercel features.
Create hooks/logger.mjs with structured logging at four levels (off, summary, debug, trace) controlled by VERCEL_PLUGIN_LOG_LEVEL env var. Support legacy VERCEL_PLUGIN_DEBUG=1 mapping to debug level.
Create src/cli/explain.ts that takes a tool name and input JSON, runs the match engine, and prints which skills would fire and why. Create src/cli/index.ts as the CLI entry point. Create src/commands/doctor.ts for diagnosing plugin setup issues.
Create src/shared/logger.ts and src/shared/match-engine.ts as TypeScript versions of the shared utilities.
Create scripts/build-manifest.ts that reads all SKILL.md files, parses frontmatter, and generates generated/skill-manifest.json.
Add CLAUDE.md documenting the architecture, dedup contract, YAML parser quirks, test conventions, and log levels.
Add tests: tests/cli-explain.test.ts, tests/logger.test.ts, tests/vercel-config.test.ts, tests/snapshot-runner.test.ts with snapshot files in tests/snapshots/. Add test fixtures in tests/fixtures/. Update existing tests to copy new hook modules (logger.mjs, vercel-config.mjs) into temp dirs.
Update package.json with new bin entry and dependencies.
The seen-skills dedup was using a temp file written via SessionStart,
but PreToolUse hooks can't update CLAUDE_ENV_FILE. This caused errors
when the temp file became stale or inaccessible across session clears.
Now VERCEL_PLUGIN_SEEN_SKILLS is a comma-delimited string stored
directly in the session environment. SessionStart initializes it as
empty, and the PreToolUse hook reads/updates process.env in-place.
Extract glob-to-regex and seen-skills file utilities into hooks/patterns.mjs
for reuse across the PreToolUse hook and validation script. Add
validateSkillMap to skill-map-frontmatter.mjs. Expand plugin tags for
chat-sdk/bot variants. Comprehensive tests for all new modules.
John Lindquist