AGENT SKILL

validate

Use when the user wants final completion sign-off on any work item — epic, story, bug, or spike: every task in docs/work/{work-id}/tasks.md verified against Gherkin AC and, for an epic, roadmap phase exit criteria. Triggers on "validate CHK01", "validate JIRA-123", "is this story done", "sign off the epic". Builds an acceptance matrix with evidence, updates task and work-item status, and produces a validation report. Do NOT use for PR or branch code review (code-review), writing tasks (tasks), sprint retrospective (sprint-retro), or drafting the breakdown (tasks) or the technical design (tdd).

License 许可
MIT
最近更新
2026-08-20
Stars
★ 2
Forks
0

01 前言 Front Matter

name
validate
license
MIT
metadata
{ "owner": "delivery", "author": "Carinya Parc", "version": "2.0", "work_shape": "review-and-gate", "output_class": "decision-support" }
description
Use when the user wants final completion sign-off on any work item — epic, story, bug, or spike: every task in docs/work/{work-id}/tasks.md verified against Gherkin AC and, for an epic, roadmap phase exit criteria. Triggers on "validate CHK01", "validate JIRA-123", "is this story done", "sign off the epic". Builds an acceptance matrix with evidence, updates task and work-item status, and produces a validation report. Do NOT use for PR or branch code review (code-review), writing tasks (tasks), sprint retrospective (sprint-retro), or drafting the breakdown (tasks) or the technical design (tdd).
allowed-tools
Read Write Edit Glob Grep Bash(git remote:*) Bash(gh:*) Bash(glab:*)
argument-hint
<work-id>
compatibility
Tracker resolution uses Linear, Atlassian (Jira), or GitHub/GitLab MCP tools when available, or `git remote`/`gh`/`glab`; falls back to the filesystem when none are reachable.

02 资源文件 3

03 SKILL.md 正文

Validate

You are a QA Lead performing a final stakeholder review to confirm a work item is production-ready and every acceptance criterion is satisfied — an epic, a story, a bug fix, or a spike's answered question.

Read work-item-resolution.md first — it resolves the source system, canonical ID, and type for {work-id} before you read anything else. Read delivery-conventions.md for paths and artefact boundaries. When the target is a story, bug, or spike, also note its parent epic — the report should say whether this item's completion unblocks or completes that epic.

Inputs

Input Location Required
Product backlog / tracker docs/product/backlog.md, or the tracker directly Yes
Tasks docs/work/{work-id}/tasks.md Yes
Design docs/work/{work-id}/tdd.md (or a legacy design.md) If exists
Application code src/ (or repo equivalent) Yes
Solution architecture docs/architecture/solution.md If relevant
ADRs docs/architecture/decisions/ If relevant

Sub-agents

When the work item has many tasks (roughly >5) or complex Gherkin, spawn ac-evidence-verifier (agents/ac-evidence-verifier.md) to build the acceptance matrix before writing the report and updating tasks.md.

For eval runs on skills in this repo, use root eval-grader (agents/eval-grader.md).

Steps

Phase 1: Gather context

  1. Locate the work item row — docs/product/backlog.md (filesystem-only) or the tracker directly (Title, type, work path docs/work/{work-id}/).
  2. Read docs/work/{work-id}/tasks.md and collect all tasks.
  3. Read docs/work/{work-id}/tdd.md if it exists, falling back to a legacy design.md.
  4. Read the solution architecture if the work item touches architectural boundaries.
  5. Read any ADRs referenced by the design or requirements.

Phase 2: Build the acceptance matrix

For every task in docs/work/{work-id}/tasks.md, build a table:

Task Criterion Evidence Status
CF-XX-YY Description of the criterion File path, test name, or observation pass / fail / partial
  • pass — criterion fully satisfied with evidence in the codebase
  • fail — criterion not met: no evidence found, or implementation contradicts it
  • partial — some aspects met but gaps remain; describe what is missing

Phase 3: Validate against code

For each acceptance criterion:

  1. Search the application codebase for the implementation.
  2. Read the relevant source files and confirm the behaviour described.
  3. Check for unit or integration tests covering the criterion.
  4. If the criterion references configuration, environment variables, or infrastructure, confirm they are present and documented.
  5. Record the evidence (file path + line range, test name, or observation).

Be thorough. Do not assume a criterion is met because a file exists — read the code and confirm the logic matches the requirement.

Phase 4: Validate against design

If a design document exists, confirm the implementation matches the specified architecture (components, data flow, interfaces), API contracts (signatures, schemas, error codes), data models, and performance and security controls. Note any deviations — they are not automatic failures, but must be documented.

Phase 5: Cross-cutting checks

Check What to verify
Tests Unit and integration tests exist and cover each public interface
Types No any casts that bypass type safety on public boundaries
Error handling Errors handled as specified in design; no silent swallows
Documentation README, runbooks, or inline docs updated if required by acceptance criteria
Environment New environment variables added to .env.example
Dependencies No unused or undeclared dependencies

Phase 6: Update tasks and backlog

Based on the acceptance matrix, update docs/work/{work-id}/tasks.md:

  1. Completed criteria — check the box - [x].
  2. Incomplete or partial criteria — uncheck the box - [ ] and append a brief note explaining what remains (e.g. — not wired to scheduler).
  3. Task status — all criteria pass → done; some fail or partial → in-progress; none pass → not started.
  4. New tasks — if validation reveals uncovered work, add tasks following existing ID and format conventions.
  5. Work item status — update in docs/product/backlog.md (filesystem-only) or the tracker only when every task for this work item is verified done. For a story/bug/spike, also note in the report whether its parent epic is now closer to done — do not update the epic's status yourself unless the user is validating the epic itself.

Phase 7: Pre-report validation

  • [ ] Work item resolved per work-item-resolution.md — asked the user on any ambiguity in source system or ID
  • [ ] Every task in docs/work/{work-id}/tasks.md appears in the acceptance matrix
  • [ ] No criterion marked pass without concrete evidence (path, test, behaviour)
  • [ ] tasks.md and backlog.md (or tracker) updates preserve existing ID and format conventions
  • [ ] Work item status set to complete only if all its tasks are verified done

Phase 8: Produce the validation report

Use the output format below.

Quality rules

  • Every acceptance criterion must be evaluated — none may be skipped
  • Evidence must be specific: cite file paths, function names, test names
  • Do not mark a criterion pass without reading the implementing code
  • Do not mark a criterion fail without searching thoroughly (multiple file patterns, grep for key terms, review related modules)
  • Deviations from the design are findings, not automatic failures — document the deviation and whether it is acceptable
  • Task and backlog updates must preserve existing format and conventions

Negative constraints

A validation report MUST NOT:

  • Write new acceptance criteria — it verifies criteria already in tasks.md
  • Include implementation detail → that belongs in solution.md or tdd.md
  • Reopen decisions closed during the sprint → raise a follow-up story instead
  • Include business rationale → that belongs in product.md
  • Judge the diff — that is code-review; validate judges work-item done-ness vs AC

Output format

<example>

Validation Report — CF-XX: Work Item Title

Date: YYYY-MM-DD Validator: AI QA Review Resolved: {source system} · {type} · {canonical ID} Work item status: complete | incomplete

Summary

{1-2 sentence summary: how many stories, how many criteria, overall result}

Acceptance Matrix

Story Criterion Evidence Status
CF-XX-01 Description path/to/file.ts L12-45 pass
CF-XX-01 Description path/to/test.ts::test name pass
CF-XX-02 Description Not found in codebase fail
CF-XX-03 Description Partially implemented in file.ts — missing X partial

Design Deviations

Area Design spec Actual implementation Assessment
API method POST /api/foo PUT /api/foo Acceptable — aligned with REST conventions

(Omit this section if no deviations found.)

Findings

  • [fail] CF-XX-02 criterion Y: {what is missing}
  • [partial] CF-XX-03 criterion Z: {what remains}
  • [observation] {any other notable finding}

(Omit this section if all criteria pass.)

Backlog Changes

  • CF-XX-01: status updated to done, all criteria checked
  • CF-XX-02: criterion Y unchecked, note added
  • CF-XX-04 (new): {title of new story added to address gap}

Conclusion

{Is this work item ready for stakeholder sign-off? If not, what must be resolved first?}

</example>

04 同仓库更多 Skills

carinyadigital
carinyadigital/skills

skill-name

Replace with description of the skill and when the agent should use it.

Skill与智能体开发
★ 2
carinyadigital
carinyadigital/skills

ux-design-review

Use when the user wants a UX or design review of implemented UI — components, pages, or flows — against its design source, covering accessibility (WCAG 2.2 AA), interaction states, responsiveness, design fidelity, and design-system conformity. Triggers on "review the checkout page", "check this UI before I ship it", "is this accessible", "does this match the design". Live-environment-first: drives the rendered UI in a real browser where possible. Works with any framework or design tooling. Produces a structured verdict with a coverage statement; writes no source changes. Do NOT use to address or fix UX findings (ux-design-fix), for code correctness or security review (code-review), to review a tdd.md document (docs-review), or to implement features (implement).

设计师视觉设计Skill与智能体开发
★ 2
carinyadigital
carinyadigital/skills

ux-design-fix

Use when the user wants to change how existing UI looks or behaves — fix spacing, colour, layout, states, responsiveness, or accessibility on a component, page, or flow. Takes either a ux-design-review output (scoped by action tier) or a direct instruction. Triggers on "the summary card padding is off", "the modal breaks on mobile", "make the focus ring visible", "fix the UX review findings", "this button looks wrong". Fixes via design tokens and library components rather than local patches, re-renders to verify visually, and re-checks neighbours for shift. Do NOT use to build new UI, screens, or flows (implement), to review UI without changing it (ux-design-review), to fix code-review findings (code-review-fix), or to redesign a component or flow.

开发工程师设计师视觉设计Skill与智能体开发
★ 2
carinyadigital
carinyadigital/skills

tdd

Use when the user wants a technical design document (TDD) for any work item — epic, story, bug, or spike — at docs/work/{work-id}/tdd.md, in skeleton or full mode. Pass a work item ID (CHK01, or a Linear/Jira/GitHub key like JIRA-123). Resolves the source system and writes at whatever level the ID names — a story gets its own tdd.md beside its parent epic's, not nested inside it. Cite solution.md — do not re-narrate architecture. Triggers on "tdd CHK01", "write the technical design for JIRA-123", "design the epic", "how should we build this story". For reviewing an existing tdd.md, use docs-review instead. Do NOT use for test-driven development — writing a failing test first, red/green/refactor, or any test-authoring task is implement. Do NOT use to write the breakdown itself — stories, tasks, or sub-tasks (tasks), task Gherkin (tasks), system-wide architecture (solution), ADR write (adr), or code implementation (implement).

视觉设计Skill与智能体开发
★ 2
carinyadigital
carinyadigital/skills

tasks

Use to decompose anything into delivery work — a product strategy and roadmap into epics, an epic into stories and tasks with Gherkin acceptance criteria, a story into sub-tasks, or an external spec, RFC, or PRD into both in one pass. Resolves the work item's source system (Linear, Jira, GitHub/GitLab issues, or filesystem) and its type first. Writes docs/product/backlog.md and docs/work/{work-id}/tasks.md, or the tracker directly when one is resolved. Triggers on "break this spec into a backlog", "write the epics", "decompose checkout-foundation", "turn this RFC into tickets", "write tasks for CHK01", "write sub-tasks for JIRA-123", "what stories do we need". EARS with --ears. Do NOT use to groom an existing backlog or check sprint readiness (backlog-refine), write technical design at any level (tdd), write architecture (solution), phase the delivery (roadmap), implement code (implement), or sign off a work item (validate).

产品经理Skill与智能体开发
★ 2
carinyadigital
carinyadigital/skills

sprint-retro

Use when the user wants to run a retrospective after a sprint ends — compare what was committed in plan.md against what actually shipped, surface themes with evidence, and record actions with owners routed to the skill that will action them, at docs/work/sprint-{id}/retrospective.md. Triggers on "run the retro", "sprint retrospective", "what went wrong last sprint", "how did sprint 3 go". Do NOT use to plan the next sprint (sprint-planning), sign off a work item against its acceptance criteria (validate), run a cross-document alignment pass (docs-review), or review a diff (code-review).

★ 2