validate-phase

<!-- markdownlint-disable MD012 MD046 -->

STOP -- DO NOT READ THIS FILE. You are already reading it. This prompt was injected into your context by Claude Code's plugin system. Using the Read tool on this SKILL.md file wastes tokens. Begin executing Step 1 immediately.

/pbr:validate-phase -- Post-Build Validation Quality Gate

References: @references/ui-brand.md

You are the orchestrator for /pbr:validate-phase. This skill identifies validation gaps in a completed phase -- requirements without tests, tests that fail, or verify commands without automated coverage -- and spawns the nyquist-auditor agent to fill them. Your job is to collect gaps, present them, delegate test generation, and write VALIDATION.md.

Context Budget

Reference: skills/shared/context-budget.md for the universal orchestrator rules. Reference: skills/shared/agent-type-resolution.md for agent type fallback when spawning Task() subagents.

Additionally for this skill:

• Minimize reading subagent output -- read only VALIDATION.md frontmatter for summaries after the auditor completes.
• Delegate all test generation to the nyquist-auditor agent. Never write tests yourself.

Step 0 -- Immediate Output

Before ANY tool calls, display this banner:

+==============================================================+
|  PLAN-BUILD-RUN > VALIDATING PHASE {N}                       |
+==============================================================+

Where {N} is the phase number from $ARGUMENTS. Then proceed to Step 1.

Multi-Session Sync

Before any phase-modifying operations (writing VALIDATION.md, spawning agents), acquire a claim:

acquireClaim(phaseDir, sessionId)

If the claim fails (another session owns this phase), display: "Another session owns this phase. Use /pbr:progress to see active claims."

On completion or error (including all exit paths), release the claim:

releaseClaim(phaseDir, sessionId)

Prerequisites

• .planning/config.json exists
• Phase has been built: SUMMARY.md files exist in .planning/phases/{NN}-{slug}/

Argument Parsing

Parse $ARGUMENTS according to skills/shared/phase-argument-parsing.md.

| Argument | Meaning | |----------|---------| | 3 | Validate phase 3 | | 3 --auto | Validate phase 3, skip interactive gap confirmation | | (no number) | Use current phase from STATE.md |

Orchestration Flow

Execute these steps in order.

Step 1: Parse and Validate (inline)

1. Parse $ARGUMENTS for phase number and --auto flag
2. If --auto is present: set auto_mode = true. Log: "Auto mode enabled -- skipping gap confirmation prompt"
3. Read .planning/config.json

CRITICAL (hook-enforced): Write .active-skill NOW. Write the text "validate" to .planning/.active-skill using the Write tool.

1. Resolve phase directory .planning/phases/{NN}-{slug}/
2. Validate:
3. Phase directory exists at .planning/phases/{NN}-{slug}/
4. SUMMARY.md files exist (phase has been built)
5. PLAN.md files exist (needed for requirement extraction)
6. If no phase number given, read current phase from .planning/STATE.md

Validation errors:

If phase directory not found:

+==============================================================+
|  ERROR                                                       |
+==============================================================+

Phase {N} not found in .planning/phases/. Run /pbr:status to see available phases.

If no SUMMARY.md files:

+==============================================================+
|  ERROR                                                       |
+==============================================================+

Phase {N} hasn't been built yet.

**To fix:** Run `/pbr:build {N}` first.

If no PLAN.md files:

+==============================================================+
|  ERROR                                                       |
+==============================================================+

Phase {N} has no plans.

**To fix:** Run `/pbr:plan-phase {N}` first.

Step 2: Collect Requirements and Tests (inline)

1. Read all PLAN-*.md files in the phase directory:
2. Extract must_haves from frontmatter (truths, artifacts, key_links)
3. Extract <verify> commands from each <task> element
4. Read all SUMMARY-*.md files in the phase directory:
5. Extract key_files from frontmatter
6. For each key_file: check if a corresponding test file exists using project test conventions:
7. .test.js, .test.ts (Jest/Vitest)
8. test_*.py (pytest)
9. *_test.go (Go)
10. .spec.js, .spec.ts (Mocha/Jasmine)
11. For each <verify> command in plans: run it, record pass/fail
12. Build gap list with entries:

{
  task_id: "string",
  requirement: "from must_have",
  gap_type: "no_test_file | test_fails | no_automated_command",
  impl_file: "path/to/file"
}

Step 3: Present Gaps (inline)

Display gap summary table:

Validation Gap Summary
| Task ID | Requirement | Gap Type | Impl File |
|---------|-------------|----------|-----------|
| T1      | {req}       | {type}   | {file}    |

If zero gaps:

1. Write VALIDATION.md with all-green status (see Step 5 format)
2. Display: "All requirements have passing tests."
3. Route to /pbr:review {N}
4. Clean up and STOP

If gaps found and --auto flag set:

Skip confirmation, proceed directly to Step 4.

If gaps found without --auto:

Use AskUserQuestion:

question: "Found {N} validation gaps. Spawn nyquist-auditor to generate tests?"
header: "Validation Gaps"
options:
  - label: "Yes"   description: "Generate tests for all gaps"
  - label: "No"    description: "Skip test generation, save gap report only"

• If "Yes": proceed to Step 4
• If "No": write VALIDATION.md with gaps (status yellow/red), display routing, clean up and STOP

Step 4: Spawn Nyquist Auditor (delegated)

Display: Spawning nyquist-auditor... (est. 2-5 min)

Build spawn prompt with:

• <files_to_read> block listing: all PLAN.md files, all SUMMARY.md files, implementation files from gaps
• <gaps> XML block with the gap list from Step 2:

<gaps>
  <gap task_id="{id}" gap_type="{type}" impl_file="{file}">
    {requirement text}
  </gap>
  ...
</gaps>

• Phase number and phase slug

Spawn the agent:

Task({
  subagent_type: "pbr:nyquist-auditor",
  prompt: <auditor prompt with gaps and files_to_read>
})

Parse agent return for completion marker:

• GAPS FILLED -- all gaps resolved
• PARTIAL -- some gaps remain
• ESCALATE -- critical issues found

Step 5: Write VALIDATION.md (inline)

CRITICAL: Write VALIDATION.md NOW. Do not skip this step.

Template reference: Read ${CLAUDE_PLUGIN_ROOT}/templates/VALIDATION.md.tmpl for the standard validation format (test infrastructure, sampling rate, per-task verification map, Wave 0 requirements).

Write .planning/phases/{NN}-{slug}/VALIDATION.md with this structure:

Frontmatter:

---
phase: {NN}
phase_slug: "{slug}"
status: "green|yellow|red"
gaps_total: {N}
gaps_resolved: {M}
gaps_escalated: {K}
---

Status values:

• green -- all gaps resolved or no gaps found
• yellow -- some gaps resolved, some remain
• red -- escalations exist or critical gaps unresolved

Body:

## Requirement-to-Test Mapping

| Task ID | Requirement | Test File | Command | Status |
|---------|-------------|-----------|---------|--------|
| T1      | {req}       | {file}    | {cmd}   | PASS   |
| T2      | {req}       | -         | -       | GAP    |

## Escalated Gaps

{If any escalated gaps, list them here with details}

{If none: "No escalated gaps."}

Step 6: Route Next (inline)

If status is green or yellow (all gaps resolved or partially resolved):

+==============================================================+
|  VALIDATION COMPLETE                                         |