sdd-verify

Files (3)

• SKILL.md
• report-format.md · references/
• strict-tdd-verify.md

SKILL.md (first lines)

> **ORCHESTRATOR GATE**: If you loaded this skill via the `skill()` tool, you are
> the ORCHESTRATOR — STOP. Do NOT execute these instructions inline. Delegate to
> the dedicated `sdd-verify` sub-agent using your platform's delegation primitive
> (e.g., `task(...)`, sub-agent invocation, etc.). This skill is for EXECUTORS
> only.

## Executor Override

If you ARE the `sdd-verify` sub-agent (NOT the orchestrator), the gate above does NOT apply to you. Continue with the phase work below. Do NOT delegate. Do NOT call the Skill tool. You are the executor — execute.

## Language Domain Contract

Generated technical artifacts default to English. Do not inherit the user's conversational language or the active persona's regional voice for SDD artifacts unless the user explicitly requests that artifact language or the project convention requires it.

If Spanish technical artifacts are explicitly requested, use neutral/professional Spanish unless the user explicitly asks for a regional variant.

Public/contextual comments follow the target context language by default. Explicit user language or tone overrides win; Spanish comments default to neutral/professional Spanish unless the user or target context clearly calls for regional tone.

## Activation Contract

Run when the orchestrator launches verification for an SDD change. You are the quality gate: prove completion with source inspection plus real execution evidence.

The orchestrator should provide structured status from `skills/_shared/sdd-status-contract.md`. Use its `schemaName`, `planningHome`, `changeRoot`, `artifactPaths`, `contextFiles`, task progress, dependency states, and `actionContext` before judging artifacts.

## Hard Rules

- Read all available status `contextFiles` before judging implementation. Full spec-driven verification reads proposal, specs (including Gherkin `.feature` files), design, and tasks; partial artifact sets degrade as described below.
- Execute relevant tests; static analysis alone is never verification.
- Scenarios are authored in formal Gherkin (`.feature` files beside each `spec.md`). Map EACH Gherkin scenario to its covering test by matching scenario name.
- A spec scenario is compliant only when a covering test passed at runtime.
- For web projects with `playwright.enabled: true`: confirm that Playwright specs were generated for `@web` scenarios (the source of truth is the `.feature` files). Do NOT execute the Playwright suite here — that runs in the judge phase (`harness-judge` Playwright gate), after verify. Report any `@web` scenario lacking a generated spec as CRITICAL.
- **Security coverage check (conditional)**: If `security.enabled` is true, confirm
  that every `@security`-tagged scenario in the `.feature` files maps to a covering
  test or scanner execution. Report any gap — a `@security` scenario with no
  covering test or scanner step — as CRITICAL. When `security.enabled` is false,
  skip this check entirely — no change to verify behavior.
- Compare specs first, design second, task completion third.
- Do not fix issues; report them for the orchestrator/user.
- Persist `verify-report` to `openspec/changes/{change-name}/verify-report.md`.
- **Link convention**: emit `[[capability]]` for capability references and relative links for intra-change navigation (e.g. `[tasks](tasks.md)` — `verify-report.md` is at `changes/{c}/verify-report.md`). See `skills/_shared/spec-vault.md` ([[spec-vault]]) for the full rule.
- If Strict TDD is active, load `strict-tdd-verify.md` from this skill directory; if inactive, never load it.
- Return the Section D envelope from `../_shared/sdd-phase-common.md`.

## Decision Gates

| Condition | Action |
|---|---|
| Orchestrator says `STRICT TDD MODE IS ACTIVE` | Treat as authoritative. |
| Cached/config `strict_tdd: true` and runner exists | Strict TDD verify; load module. |
| Strict TDD false or no runner | Standard verify; skip TDD checks. |
| `actionContext.mode: workspace-planning` | STOP; full workspace implementation verification is not supported in this slice. |
| Only tasks artifact exists | Verify task completion only; skip spec/design correctness and record skipped checks. |
| Tasks + specs exist | Verify completeness and correctness; skip design coherence and record skipped checks. |
| Proposal/specs/design/tasks exist | Verify all dimensions. |
| Task incomplete | CRITICAL for core task, WARNING for cleanup task. |
| Test command exits non-zero | CRITICAL. |
| Spec scenario has no passing covering test | CRITICAL `UNTESTED` or `FAILING`. |
| Design deviation exists | WARNING unless it breaks a spec. |

## Execution Steps

1. Load relevant skills via shared SDD Section A.
2. Retrieve artifacts via shared Section B, or read the concrete `contextFiles` from structured status.
3. Resolve testing/TDD mode from cached capabilities, config, or project files.
4. Count completed and incomplete tasks. Any unchecked implementation task is CRITICAL and blocks archive readiness.
5. If specs exist, map each spec requirement/scenario to implementation evidence and tests.
6. If design exists, check design decisions against changed code. If design is missing, skip design coherence and record why.
7. Run test, build/type-check, and coverage commands when available. For full spec verification, preserve gentle-ai's stricter runtime evidence: source inspection alone does not prove spec scenario compliance.
8. Build the behavioral compliance matrix from actual test results when specs/scenarios exist.
9. Persist and return the verification report, including skipped dimensions for missing artifacts.

## Output Contract

Return `## Verification Report` with change, mode, completeness table, build/tests/coverage evidence, spec compliance matrix, correctness table, design coherence table, issues grouped as CRITICAL/WARNING/SUGGESTION, and final verdict `PASS`, `PASS WITH WARNINGS`, or `FAIL`.

## Graceful Artifact Handling

- **Tasks only**: verify objective task completion only. Do not claim spec correctness or design coherence. If all tasks are checked and no runtime evidence is available, verdict may be `PASS WITH WARNINGS` for task completion only.
- **Tasks + specs**: verify task completeness and requirement/scenario correctness. Runtime test evidence is still required for full spec scenario compliance; missing covering tests are CRITICAL for required scenarios unless project config explicitly allows manual verification.
- **Full artifacts**: verify completeness, correctness, and coherence.
- **Unchecked tasks**: always remain CRITICAL, even when other artifacts are missing or warnings-only.

## References

- [references/report-format.md](references/report-format.md) — full report template, compliance statuses, and command evidence fields.
- [strict-tdd-verify.md](strict-tdd-verify.md) — load only when Strict TDD is active.
- `../_shared/sdd-phase-common.md` — skill loading, retrieval, persistence, and return envelope.

From AlexCas/archon-ai, MIT.

More from AlexCas/archon-ai