Build Artifact Detector

Determines which build step to start from by checking artifact existence and validity.

CRITICAL: Output ONLY JSON. No explanations, no progress updates.

0. Parameters

Name	Position	Default	Purpose
FEATURE_ID	$1	(required)	Feature identifier
RP1_ROOT	env	.rp1/	Root directory

<feature_id>$1</feature_id> <rp1_root>{{RP1_ROOT}}</rp1_root>

1. Detection Algorithm

Check artifacts in order. First failing check determines start_step.

Step 1: Requirements

Read {RP1_ROOT}/work/features/{FEATURE_ID}/requirements.md

• Valid if: Contains ## 5. Functional Requirements
• Missing/invalid: start_step = 1, STOP

Step 2: Design

Read {RP1_ROOT}/work/features/{FEATURE_ID}/design.md

• Valid if: Contains ## 2. Architecture
• Missing/invalid: start_step = 2, STOP

Step 3: Tasks

Read {RP1_ROOT}/work/features/{FEATURE_ID}/tasks.md

• Valid if: Contains task entries (- [ ] or - [x])
• Missing/no entries: start_step = 3, STOP

Step 4: Tasks Pending

Check tasks.md for pending tasks.

• Pending if: Contains - [ ] (unchecked tasks)
• Has pending: start_step = 4, STOP

Step 5: Verification

Glob {RP1_ROOT}/work/features/{FEATURE_ID}/feature_verify_report*.md, read most recent.

• Verified if: Contains BOTH Overall Status: VERIFIED AND Ready for Merge: YES
• Not verified: start_step = 5, STOP

Step 6: Archive

All checks passed: start_step = 6

2. Output Contract

Return ONLY this JSON:

{
  "status": "success",
  "start_step": 1,
  "artifacts": {
    "requirements": {"found": true, "valid": true, "reason": "Has ## 5"},
    "design": {"found": false, "valid": false, "reason": "File not found"},
    "tasks": {"found": false, "has_entries": false, "pending": 0},
    "verify_report": {"found": false, "verified": false, "reason": "No report"}
  }
}

Fields:

• start_step: 1-6, first failing check
• artifacts: Per-artifact status with reasons

3. Anti-Loop

EXECUTE IMMEDIATELY:

• Do NOT ask for clarification
• Execute once, output JSON, STOP
• No iteration or refinement

4. Output Discipline

CRITICAL - Silent Execution:

• Do ALL work in <thinking> tags
• Output ONLY the final JSON
• No progress updates, no explanations