agentleFS
hushh-labshushh-labs/hushh-research/.agents/skills/quality-contracts/SKILL.md

quality-contracts

Skill40 starsChanged 4 days agoApache-2.0View on GitHub

Use when changing cross-surface verification policy, contract-test placement, test selection, or quality gates across frontend and backend.

The file

We split this skill into the rules it gives an agent

10 rules. About 0 are common, 7 are distinctive, and 3 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈7

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 3

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Very specific:

    Avoid overlap with streaming-contracts and repo-operations.

    line 12
  2. Distinctive:

    Test selection and contract-test placement decisions.

    line 33
  3. Distinctive:

    Start from the contract or user-visible behavior that needs proof.

    line 53
  4. Distinctive:

    Keep frontend and backend contract tests aligned with the same user-visible or policy-visible rule.

    line 55
  5. Distinctive:

    Do not add one-time, manual, skipped-by-default, or source-string tests to routine suites when a reusable runtime helper, component test, …

    line 56
  6. Very specific:

    Use Playwright only when a real browser is required.

    line 59
  7. Very specific:

    Treat CI pipeline ownership as repo-operations unless the primary question is what should be verified.

    line 61
  8. Distinctive:

    When changing required test sets or gate policy, prove the changed contract through its canonical entrypoint.

    line 63
  9. Distinctive:

    Treat helper-only drift as advisory unless it weakens runtime, deploy, or test authority.

    line 64
  10. Distinctive:

    Use focused or release validation for the changed contract and adjacent callers.

    line 75

≈ “Common” is an estimate: at least 1 in 5 of the 372,429 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Hussh Quality Contracts Skill

Purpose and Trigger

  • Primary scope: quality-contracts
  • Trigger on contract-test placement, test selection, route/browser verification, and cross-surface quality rules.
  • Avoid overlap with streaming-contracts and repo-operations.

Coverage and Ownership

  • Role: spoke
  • Owner family: security-audit

Owned repo surfaces:

  1. docs/development/quality
  2. hushh-webapp/__tests__
  3. consent-protocol/tests

Non-owned surfaces:

  1. security-audit
  2. frontend
  3. backend

Do Use

  1. Test selection and contract-test placement decisions.
  2. Cross-surface verification policy and quality-gate ownership.
  3. Reviewing whether a change is missing authoritative checks.

Do Not Use

  1. Broad security intake where the correct spoke is unclear.
  2. Repo-wide CI/deploy operations work.
  3. Narrow streaming-protocol work when the issue is clearly streaming only.

Read First

  1. docs/development/quality/README.md
  2. docs/development/quality/pr-impact-checklist.md
  3. docs/architecture/finance/kai-runtime-smoke-checklist.md
  4. .agents/skills/quality-contracts/references/browser-verification-contract.md
  5. docs/development/quality/definition-of-done.md

Workflow

  1. Start from the contract or user-visible behavior that needs proof.
  2. Select the smallest authoritative checks across frontend, backend, route, data, and browser surfaces.
  3. Keep frontend and backend contract tests aligned with the same user-visible or policy-visible rule.
  4. Do not add one-time, manual, skipped-by-default, or source-string tests to routine suites when a reusable runtime helper, component test, service test, or script-first smoke proves the behavior.
  5. Source-reading contract tests are allowed only for explicit architecture boundaries with no cheaper runtime surface; delete them when behavior-level coverage becomes available.
  6. Legacy, migration, and compatibility tests stay only while they protect a live compatibility/removal boundary; otherwise delete or move the durable fact into the canonical contract.
  7. Use Playwright only when a real browser is required; follow browser-verification-contract.md for protected routes.
  8. For signed-in, vault, PKM, Email Helper, consent, or app-review proof, resolve the reviewer from runtime env (REVIEWER_UID; deprecated aliases only as migration fallbacks) and assert the browser session, vault owner, workflow owner, and PKM data owner are the same user. Do not validate against copied recipients, counterparty labels, admin fixtures, or a hardcoded UID/email.
  9. Treat CI pipeline ownership as repo-operations unless the primary question is what should be verified.
  10. For new tables or data-contract changes, include the repo data-model audit.
  11. When changing required test sets or gate policy, prove the changed contract through its canonical entrypoint; do not duplicate the same selected identities in a separate direct run. Qualify the completed combined candidate once with retained functional/core checks plus affected boundaries; reuse unchanged evidence.
  12. Treat helper-only drift as advisory unless it weakens runtime, deploy, or test authority.

Handoff Rules

  1. Broad or ambiguous security work routes back to security-audit.
  2. CI or deploy pipeline ownership routes to repo-operations.
  3. Streaming-specific contract work routes to streaming-contracts.
  4. Pure frontend or backend implementation routes to frontend or backend.

Required Checks

Use focused or release validation for the changed contract and adjacent callers. Diagnostic runs the historical full set. Missing change metadata stays conservative. Explicit browser diagnostics own browser proof, and native changes require native verification. Keep live security, consent, schema and core deployment gates blocking. Secondary feature diagnostics report their actual outcome separately; no diagnostic run is not a pass. Follow the CI readiness policy.

cd hushh-webapp && npm run test:ci
cd hushh-webapp && npm run verify:service-boundary
cd consent-protocol && python3 -m pytest tests/quality -q
./bin/hushh agents data-model-audit

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.