fix(factory): complete fail-closed receipt shape validation

[groupthinking]

Canonical issue

Fixes #2139.

Outcome

Completes the fail-closed runtime-shape boundary for the fixture-only ChatGPT MCP Skills receipt. Malformed nested wire objects now produce deterministic ChatGPT fixture handoff held: ... errors instead of raw JavaScript TypeError exceptions.

Scope

• Adds a reusable record/object guard.
• Validates root input, capabilities, extension metadata, result, skill, frontmatter, resource contents, approval records, and every manifest entry before property access.
• Adds a malformed-input matrix covering null, primitive, array, and nested-object failures.
• Proves every negative case uses the controlled hold prefix and is not a TypeError.
• Preserves the existing receipt schema, hash behavior, fixture-only mode, and zero-authority boundary.

Evidence

• Residual main-branch behavior reviewed after merged fix: harden ChatGPT fixture receipt parsing #1956.
• Source merge inspected at b136aafd882440739c91ea25379e95820fc76773.
• Branch is two commits ahead of current main and touches only the parser plus its focused test.

Verification plan

• Focused Vitest receipt suite.
• Frontend type-check, lint, unit tests, and production build.
• Coverage, Security/Trivy, CodeQL, dependency review, and secret scan.
• Report preview-dependent E2E separately from the fixture-only acceptance claim.

Boundaries

No live ChatGPT import, network MCP call, authorization grant, deployment, or production activation. Keep this PR draft until exact-head checks are inspected.

Risk and rollback

Low, fail-closed input-hardening change. Roll back by reverting the two branch commits.

[vercel]

The latest updates on your projects. Learn more about Vercel for GitHub.

Project	Deployment	Actions	Updated
v0-uvai	Ready	Preview, v0	Sep 19, 2026 7:00am UTC

[coderabbitai]

Important

Draft PR not reviewed

Draft PRs are not automatically reviewed by default.

• Trigger a manual review

To automatically review draft PRs, update your CodeRabbit configuration:

reviews:
  auto_review:
    drafts: true

---

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

_{Comment @coderabbitai help to get the list of available commands.}

[github-actions]

Dependency Review

✅ No vulnerabilities or license issues or OpenSSF Scorecard issues found.

Scanned Files

None

[github-actions]

✅ E2E Test Results: ALL TESTS PASSED

Metric	Value
Status	🟢 GREEN
Total Tests
Passed	19
Failed
Deployment	https://v0-uvai-dwnm60f2k-garv1.vercel.app

Test Output

dout�[2m | tests/e2e/pipeline.test.ts�[2m > �[22m�[2mEventRelay E2E — Live Deployment�[2m > �[22m�[2mSSE Pipeline Stream�[2m > �[22m�[2mSSE stream emits at least a pipeline_status:running event
�[22m�[39m[E2E] last pipeline_status: complete

 �[32m✓�[39m tests/e2e/pipeline.test.ts�[2m > �[22mEventRelay E2E — Live Deployment�[2m > �[22mSSE Pipeline Stream�[2m > �[22mSSE stream emits at least a pipeline_status:running event�[33m 19999�[2mms�[22m�[39m
 �[32m✓�[39m tests/e2e/pipeline.test.ts�[2m > �[22mEventRelay E2E — Live Deployment�[2m > �[22mSSE Pipeline Stream�[2m > �[22mSSE stream closes within 90 seconds (no 95% hang)�[33m 27096�[2mms�[22m�[39m
 �[32m✓�[39m tests/e2e/pipeline.test.ts�[2m > �[22mEventRelay E2E — Live Deployment�[2m > �[22mSSE Pipeline Stream�[2m > �[22mSSE events fire in correct agent order�[33m 20321�[2mms�[22m�[39m
 �[32m✓�[39m tests/e2e/pipeline.test.ts�[2m > �[22mEventRelay E2E — Live Deployment�[2m > �[22mCloudEvent Schema�[2m > �[22mSSE events contain valid timestamps�[33m 24082�[2mms�[22m�[39m
 �[32m✓�[39m tests/e2e/pipeline.test.ts�[2m > �[22mEventRelay E2E — Live Deployment�[2m > �[22mCloudEvent Schema�[2m > �[22mterminal pipeline_status includes duration and stage progress�[33m 22522�[2mms�[22m�[39m
 �[32m✓�[39m tests/e2e/pipeline.test.ts�[2m > �[22mEventRelay E2E — Live Deployment�[2m > �[22mError Handling�[2m > �[22mmissing URL returns 400, not a hang�[32m 86�[2mms�[22m�[39m
 �[32m✓�[39m tests/e2e/pipeline.test.ts�[2m > �[22mEventRelay E2E — Live Deployment�[2m > �[22mError Handling�[2m > �[22minvalid URL returns error event or completes quickly, not a hang�[33m 12169�[2mms�[22m�[39m
 �[32m✓�[39m tests/e2e/pipeline.test.ts�[2m > �[22mEventRelay E2E — Live Deployment�[2m > �[22mDashboard�[2m > �[22m/dashboard returns 200�[33m 507�[2mms�[22m�[39m
 �[32m✓�[39m tests/e2e/pipeline.test.ts�[2m > �[22mEventRelay E2E — Live Deployment�[2m > �[22mDashboard�[2m > �[22m/dashboard contains agent or pipeline visualization markup�[33m 1342�[2mms�[22m�[39m
 �[32m✓�[39m tests/e2e/pipeline.test.ts�[2m > �[22mEventRelay E2E — Live Deployment�[2m > �[22mAPI Health�[2m > �[22mGET /api returns a response (not 404)�[32m 256�[2mms�[22m�[39m
 �[32m✓�[39m tests/e2e/pipeline.test.ts�[2m > �[22mEventRelay E2E — Live Deployment�[2m > �[22mAPI Health�[2m > �[22mPOST /api/pipeline/stream with no body returns 400�[32m 81�[2mms�[22m�[39m
 �[32m✓�[39m tests/e2e/pipeline.test.ts�[2m > �[22mEventRelay E2E — Live Deployment�[2m > �[22mStatic Assets�[2m > �[22mhomepage has proper meta tags�[32m 85�[2mms�[22m�[39m
 �[32m✓�[39m tests/e2e/pipeline.test.ts�[2m > �[22mEventRelay E2E — Live Deployment�[2m > �[22mStatic Assets�[2m > �[22m/features page returns 200�[33m 351�[2mms�[22m�[39m

�[2m Test Files �[22m �[1m�[32m1 passed�[39m�[22m�[90m (1)�[39m
�[2m      Tests �[22m �[1m�[32m19 passed�[39m�[22m�[90m (19)�[39m
�[2m   Start at �[22m 07:00:53
�[2m   Duration �[22m 131.47s�[2m (transform 59ms, setup 0ms, import 75ms, tests 131.27s, environment 0ms)�[22m

[groupthinking]

Execution receipt — exact head 16dd941f750f74c3255d1336851d802ab67ba745

Decision: TEST — keep draft and unmerged.

Observed correction

• Added runtime record guards for every root and nested untrusted wire-object boundary.
• Validated every manifest entry before URI, digest, or size access.
• Added a malformed-input matrix proving controlled ChatGPT fixture handoff held: failures rather than raw TypeError.
• Receipt schema, deterministic hashing, fixture-only behavior, and zero-authority policy remain unchanged.

Addressed failure

The initial head 437f93cc23ae81a63c06916e0ff008a89de02571 failed type-check and tests because the resource-content guard accidentally referenced its own block-scoped variable. This was corrected in 16dd941f. The failed run remains visible and is not reclassified.

Prevention: after mechanical identifier replacement, re-fetch and inspect the exact declaration and run type-check before treating the mutation as viable.

Verification

All applicable exact-head gates completed successfully:

• CI run 4610 — type-check, lint, frontend unit tests, production web build/image, guards, and Python matrix passed
• Coverage run 3939 — passed
• Security/Trivy run 4639 — passed
• CodeQL run 4673 — passed
• Dependency review run 4009 — passed
• Secret scan run 4023 — passed
• E2E run 4117 — passed
• No unresolved review threads were present when inspected.
• PR was open, draft, and mergeable at the verified head.

Boundary

No merge, production deployment, live ChatGPT import, external MCP execution, or authorization grant was performed.

Next action: human review of the exact head; merge requires explicit approval for PR #2140 or this exact SHA.