Skip to content
Skills Library
Programming · Review & audit

gate

/gate

Checks every acceptance criterion against executed evidence and writes a report in docs/gates/ with an APPROVE/REJECT recommendation. Use when a level-3 phase or milestone ends or the user asks if something is ready to release.

Repository
cardume ↗
Licence
MIT
Last repository change
04.11.2026

Assessment and security

100/ 100Class: good
Format25 / 25
Frontmatter with name and description, valid name, description up to 1,024 characters, name matching the folder.
Discoverability25 / 25
Description explains when to use the skill, has a reasonable length and uses third person.
Content35 / 35
At least 150 words and at most 500 lines, headings, step lists, examples or code.
Hygiene15 / 15
No placeholders such as TODO, no Windows paths, with a licence.
Riskno signals detected in the scan
Files in this package
1
Scripts
0
HTML comments
1
URLs in SKILL.md
0
Deprecation

No signals detected. Repository changed on 04.11.2026, without references to retired models or APIs.

Licence

MIT, open: use and modify while retaining author attribution.

What to check before installing
  • SKILL.md zawiera komentarze HTML (1). Nie widać ich w podglądzie Markdown, a model je czyta.

What it does and when to use it

Checks every acceptance criterion against executed evidence and writes a report in docs/gates/ with an APPROVE/REJECT recommendation. Use when a level-3 phase or milestone ends or the user asks if something is ready to release.

Original author description

Checks every acceptance criterion against executed evidence and writes a report in docs/gates/ with an APPROVE/REJECT recommendation. Use when a level-3 phase or milestone ends or the user asks if something is ready to release.

How it works

  1. 1. Find the criteria
  2. 2. Verify
  3. 3. Write docs/gates/GATE-<id>.md
  4. 4. Hand over
Original SKILL.md headings · 72 lines, 579 words
SKILL.md preview
---
name: gate
description: "Checks every acceptance criterion against executed evidence and writes a report in docs/gates/ with an APPROVE/REJECT recommendation. Use when a level-3 phase or milestone ends or the user asks if something is ready to release."
---

# Gate — evidence-based approval of a phase

The gate is the checkpoint between "built" and "approved/released". The agent prepares the decision;
**the user decides**. Gates are required at the end of a level-3 phase or milestone; at level 2 they
are optional, run on request.

**Who runs the checks.** When the main session runs this skill, it delegates steps 2–3 to the
`qa-gatekeeper` agent, so the verifier is not the one who built the work. When `qa-gatekeeper` runs
it (the skill is preloaded there), it executes the checks directly — it does not delegate to itself.

## 1. Find the criteria

- Level 3: the PRD for the phase (`docs/prd/`) and the specs under it.
- Level 2: the feature's `docs/specs/<slug>/spec.md` requirements and its `verify.md` if it exists.
- Principles: the `## Principles (manual)` section of `docs/standards.md`, plus the rest of that file
  as conventions. Do not hunt for other principle files; use a project `SPEC.md` or similar only when
  the user points to it.
- Drift: the `## Drift` section of each `verify.md` and the drift findings of the last `review`
  (`review` skill, step 4b). Open drift — code that no longer matches its approved spec — is a ❌,
  because the gate would otherwise approve criteria that describe different software.
- Gate id: `PHASE-N` for phased projects, otherwise the spec slug or version (`GATE-v1.2`).

If there is no PRD or spec, say so and stop: a gate without written criteria is an opinion.

## 2. Verify

The main session gives `qa-gatekeeper` the criteria file(s) and the scope (branch, PR list,
version). The gatekeeper runs the checks itself — tests, validators, builds, scripts,
`cardume-checklist .` — and pastes the relevant output. It reads `verify.md` as input but does not
edit it or any spec status; those belong to the `verify` skill.

Evidence rules:
- Evidence is a command with its output, a test name with its result, or a link to a file/recording in
  the repo. A statement without one of these is not evidence.
- Each criterion is ✅ (evidence shows it passes), ❌ (evidence shows it fails) or ⚠️ (could not be
  verified here — say what is needed).
- There is no "approved with caveats": any ❌ or unresolved ⚠️ means REJECT, with the list of what is missing.

## 3. Write `docs/gates/GATE-<id>.md`

```markdown
# Gate <id> — <phase/feature name>

Date: <YYYY-MM-DD> · Scope: <branch / PRs / version> · Verified by: qa-gatekeeper

## Criteria
| ID | Criterion | Result | Evidence |
|----|-----------|--------|----------|
| C1 | ... | ✅ | `pnpm test -- invoice` → 12 passed |

## Standards and principles
- <each item of Principles (manual) and each standards section that applies> — ✅/❌ — how it was checked

## Spec drift
- <spec> — in sync ✅ / drift ❌ (<what differs>)

## Recommendation
APPROVE | REJECT — <one paragraph; for REJECT, the exact list of what is missing>

## Decision
<!-- Left blank for the user: APPROVED or REJECTED, date, notes. -->
```

## 4. Hand over

Show the user the criteria table and the recommendation, and ask for the decision (or, as a
subagent, return the recommendation and the decision request under **Open questions**; the caller
asks the user). The main session writes the user's answer under `## Decision`, only when they give it. After an approved
gate, the next steps are the next phase's PRD or a release: the `release-manager` agent (or the user
runs `/cardume-core:release`).

Versions

RepositoryScoreLicence
This page · jeffersonvalente/cardume100MIT
sh5623/e2e-rail100MIT
skillberry-ai/cap-evolve100Apache-2.0
bitIO/wheelhouse100MIT
Perchusha/pitwall100MIT
jease0502/research-agent-kit100Apache-2.0
Vibecheck-QA/vibecheck100MIT
iankiku/forwward-teams100MIT
fdu-ins/insurance-skills100Apache-2.0

Database snapshot: 13.11.2026 · package version 6ef412804802 (SHA-256 digest of the entire package)

Downloads in the last 30 days: 0 · ranking

Report a problem with this skill: harmful content, incorrect descriptions or an author’s removal request.

Similar skills

  • 90protocolsutra-release-gate
    Determines readiness according to executed criteria, regression evidence and explicit release policy. Use for ProtocolSutra AI scoped interoperability analysis; distinguish executed evidence from proposed tests.
  • 90run-tests
    Use when a ticket or acceptance criterion needs the test suite run and its result evidenced, or after making a change to confirm nothing broke. Invoke for "tests must pass", "verify the suite is green", or before handing any code change to review — and to check that each…
  • 100execution-grounded-review
    Execution-grounded review: run tests first, trace each acceptance criterion to execution evidence. Use when verifying an implementation meets spec.
  • 100review-ticket
    Use as a reviewer agent to review another agent's `in_review` ticket — never your own. Judge whether each acceptance criterion is genuinely met and proven by a test that exercises that criterion's own behaviour, and whether the change is sound, then record an ADVISORY verdict…
  • 90agent-qa-gate
    Place a QA agent as the last blocking node before output ships, checking acceptance criteria against executed test evidence with the power to veto. Use when builder agents keep declaring done on work that does not meet its criteria and you need one gate that runs the artifact…
  • 100harness-accept
    Generate an Acceptance Demo HTML for non-engineer vibecoders right before ship/wait/reject decision. Reads back the acceptance_criteria that were stored as personal-preference.v1 by harness-plan-brief (joined by user_request_hash), then renders a single-file HTML showing each…

Feedback and reports

Your message goes to Bartosz Sokoliński. An email address is needed only for a reply. See the privacy policy.