Verify acceptance criteria, quality, and readiness before closure.
Install
npx skillscat add kieutrongthien/skills-fullstack/plan-task-completion-verification Install via the SkillsCat registry.
SKILL.md
Task Completion Verification Skill
What it does
- Confirms acceptance criteria are met with evidence.
- Checks quality gates (tests, QA, accessibility/performance where relevant).
- Ensures documentation and stakeholders are updated before closure.
When to use
- Before demo/release to ensure readiness.
- When handing off to QA/UAT or stakeholders.
Key operations
- Identify active project and task; review acceptance criteria and task state (
tasks/*.json). - Verify implementation against criteria; capture evidence (screenshots, logs, links).
- Validate tests (unit/integration/e2e) and linters; check performance/a11y/security as applicable.
- Update docs/changelog/release notes; confirm stakeholder sign-off or UAT.
- Move task state to testing/done per
AGENTS.md, log results and time, updatememory/YYYY-MM-DD.md.
Output files
projects/<ACTIVE>/tasks/in-progress.json→tasks/done.json(state transition when complete).projects/<ACTIVE>/tasks/task-log.md— evidence, test results, time spent, blockers resolved.projects/<ACTIVE>/verification/checklist.md— criteria, evidence, test results, sign-offs (if repo uses verification/).projects/<ACTIVE>/verification/test-results.md— summary of automated/manual test outcomes.projects/<ACTIVE>/verification/release-notes.md— user-facing changes and impacts (if shipping).memory/YYYY-MM-DD.md— decisions, sign-offs, risks.
Example workflow
- Identify active project and task.
- Re-read acceptance criteria; walk through implementation against each with evidence.
- Run automated tests; execute targeted manual checks (a11y/perf/security if needed); record in
test-results.md/task-log.md. - Update
checklist.mdandrelease-notes.md; obtain sign-off (QA/PO); move task todone.json; log inmemory/YYYY-MM-DD.md.
Best practices
- Evidence-first: link PRs, screenshots, logs for each criterion.
- Do not skip tests; rerun after late changes; note flaky cases.
- Include a11y/perf spot checks for UI; include security considerations for APIs.
- Keep checklists short, consistent, and repeatable.
- Update task state immediately after verification; keep audit trail in
task-log.mdandmemory/YYYY-MM-DD.md. - Avoid deleting/overwriting artifacts; follow
AGENTS.mdsafety (no destructive operations).
Related resources
AGENTS.md(task flow, safety rules)SKILLS.md,MEMORY.md,memory/YYYY-MM-DD.mdprojects/<ACTIVE>/tasks/*.json,task-log.md- Definition of Done; testing/coverage reports; a11y/perf playbooks (team docs)