test(#12979): cover legacy A2A paid skill dispatch

[lalalune]

Summary

• Add legacy message/send coverage for the disabled paid A2A skill ids: chat_with_agent, video_generation, and generate_video.
• Add the required issue evidence artifact for a2a skills.ts/handlers.ts: latent money hazards in EXPORTED-but-unwired skill dispatch (guard before wiring) #12979 after fix(#12979): fail closed unwired A2A paid skills #13020 merged the fail-closed guard.

Follow-up to #13020 / #12979.

Evidence

• .github/issue-evidence/12979-a2a-skill-money-guards.md
• bun test --isolate --coverage-reporter=lcov packages/cloud/shared/src/lib/api/a2a/skills.money-guard.test.ts — PASS, 4 tests, 13 assertions
• bun run --cwd packages/cloud/shared typecheck — PASS
• bun run --cwd packages/cloud/shared lint — PASS
• git diff --check origin/develop..HEAD — PASS

Evidence N/A

• UI screenshots/video/frontend logs: N/A — backend-only fail-closed guard test/evidence follow-up.
• Real LLM trajectories: N/A — no model path is invoked; the tests prove disabled paths do not reach model/provider services.
• DB/billing artifacts: N/A — expected behavior is absence of credit reservations, generation rows, and agent dispatch.

[greptile-apps]

Your trial has ended. Reactivate Greptile to resume code reviews.

[coderabbitai]

Important

Review skipped

Auto reviews are disabled on this repository. Please check the settings in the CodeRabbit UI or the .coderabbit.yaml file in this repository. To trigger a single review, invoke the @coderabbitai review command.

⚙️ Run configuration

Configuration used: Organization UI

Review profile: CHILL

Plan: Pro

Run ID: 29f2c9d8-772f-4d96-abfb-5b2cc77e7df6

You can disable this status message by setting the reviews.review_status to false in the CodeRabbit configuration file.

Use the checkbox below for a quick retry:

• 🔍 Trigger review

✨ Finishing Touches

🧪 Generate unit tests (beta)

• Create PR with unit tests
• Commit unit tests in branch test/12979-a2a-dispatch-evidence

---

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.}

[lalalune]

Validated locally on pr-13025 (ffb2ec9b4b).

Passed:

• bun test --isolate --coverage-reporter=lcov packages/cloud/shared/src/lib/api/a2a/skills.money-guard.test.ts — 4 pass / 0 fail / 13 expects
• bun run --cwd packages/cloud/shared lint
• bun run --cwd packages/cloud/shared typecheck after bun run build:core repaired the local workspace dist graph
• git diff --check github/develop...HEAD

Manual review notes:

• The new legacy message/send loop covers chat_with_agent, video_generation, and generate_video and asserts they reach the disabled paid-skill guard.
• The allowed task/history/moderation mocks are scoped to the handler path; paid/model/provider mocks still throw if reached.
• Evidence file correctly marks runtime artifacts N/A for an absence-of-side-effects backend guard.

I did not merge because GitHub currently reports merge state UNKNOWN; the local diff and focused validation look good.

[lalalune]

Reviewed: test + evidence-doc only, no production code. Covers legacy A2A paid-skill dispatch via the money-guard. (Minor note: the mock relaxation from throwing-sentinels to no-op mocks slightly loosens the 'must not be called' strictness — fine for this coverage add.) Queuing auto-merge on green.