Skip to content

test(app): add learning evaluation seam typed coverage matrix - #95

Merged
riquelmechile merged 1 commit into
mainfrom
feat/learning-app-seam-coverage
Aug 21, 2026
Merged

riquelmechile merged 1 commit into
mainfrom
feat/learning-app-seam-coverage

Conversation

@riquelmechile

Copy link
Copy Markdown
Owner

Linked issue

Closes #89

PR type

  • Bug fix (type:bug)
  • New feature (type:feature)
  • Documentation only (type:docs)
  • Code refactoring (type:refactor)
  • Maintenance / tooling (type:chore)
  • Breaking change (type:breaking-change)

Summary

  • Work unit 3 child W3C3 (auto-split, stacked-to-main on merged W3C2 test(app): add learning evaluation seam gate guards #94, 340 ≤ 400 lines): extended typed matrix for the evaluation seam (task 2.2 completion) — test-only child, zero production lines.
  • 9 table-driven families over the merged seam: typed authority failures never promote (authority-missing / authority-malformed / authority-foreign via it.each; unresolvable sourceRef; revoked superseding proof → authority-revoked); skill-not-found / skill-not-active fail closed before policy work; malformed explicit evidence + zero outcomes typed and never harmful; loaded unresolved conflict escalates to typed needs-review; foreign-tenant events excluded even when the stream leaks them (outcomeIds stay tenant-scoped).
  • Tasks 2.1 + 2.2 now complete. Full gate pnpm check exit 0 — 1576 passed | 6 skipped, 0 new warnings.

Changes

File or area Change
packages/app/test/learning/evaluate-coverage.test.ts Typed authority-failure matrix + guards/wiring tests (new, 339)
openspec/changes/learning-promotion/tasks.md Task 2.2 checked

Verification

  • Focused checks pass: pnpm exec vitest run packages/app/test/learning/evaluate-coverage.test.ts → 9/9
  • The affected runtime scenario was exercised, or is not applicable (test-only child over the pure seam)
  • Full gate: pnpm check GREEN — 1576 passed | 6 skipped, 0 new warnings
  • Native RDD: lineage review-8e7d436130116786 (medium, reliability lens) approved with focused+full-gate evidence; pre-pr gate allow

Contributor checklist

  • The linked issue is open and has status:approved.
  • Exactly one PR type is selected and its matching type:* label is applied.
  • The branch name follows type/lowercase-description.
  • Tests and documentation are updated where behavior changed.
  • Commits use Conventional Commit format and contain no attribution trailers.
  • No secrets or unrelated changes are included.

Work unit 3 child 6 (learning-promotion): typed authority failures never
promote, skill misresolution guards, malformed explicit evidence, zero
outcomes, unresolved-conflict escalation, and foreign-event isolation.
@riquelmechile riquelmechile added the type:feature Feature pull request label Aug 21, 2026
@riquelmechile
riquelmechile merged commit 546f3ba into main Aug 21, 2026
2 of 3 checks passed
@riquelmechile
riquelmechile deleted the feat/learning-app-seam-coverage branch August 21, 2026 23:49

@chatgpt-codex-connector chatgpt-codex-connector Bot left a comment

Copy link
Copy Markdown

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: 4e03b005e4

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

  • Open a pull request for review
  • Mark a draft as ready
  • Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

Comment on lines +165 to +168
it.each([
['authority-missing', undefined],
['authority-malformed', { companyId: 'acme', subject, extra: 1 }],
['authority-foreign', { companyId: 'other', subject, sourceRef: 'proof-1' }],

Copy link
Copy Markdown

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

P2 Badge Expand the authority failure matrix before closing task 2.2

This matrix and the two cases below exercise only authority-missing, authority-malformed, authority-foreign, and authority-revoked, while design.md:48,66 requires coverage of every authority-unavailable reason. The app seam therefore remains untested for authority-stale, authority-ambiguous, authority-source-unresolvable, command/principal/policy mismatches, and authority-proof-unavailable, even though tasks.md now reports the typed matrix as complete; regressions that rewrite or fail to propagate any of those repository results would pass this suite.

Useful? React with 👍 / 👎.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

type:feature Feature pull request

Projects

None yet

Development

Successfully merging this pull request may close these issues.

[Feature]: Wire promotion evaluation app seam, fakes, and atomic verify hook

1 participant