| name | implementation-readiness |
| description | Gate for HIGH/EPIC tasks that validates plan completeness and architecture compliance before implementation begins |
| version | 1 |
| model | sonnet |
| invoked_by | both |
| user_invocable | true |
| tools | ["Read","Glob","Grep"] |
| verified | true |
| best_practices | ["Run before spawning implementation agents for HIGH/EPIC tasks","Skip for TRIVIAL/LOW complexity tasks","Produce structured pass/fail verdict with reasons"] |
| error_handling | graceful |
| streaming | supported |
| source | builtin |
| trust_score | 100 |
| provenance_sha | 3b709cb6f8c80bf6 |
Implementation Readiness Gate
Validates that a plan is ready for implementation before agents are spawned. Prevents wasted compute from incomplete plans, missing acceptance criteria, or architectural misalignment.
When to Invoke
Skill({ skill: 'implementation-readiness' });
Invoke when:
- Router is about to spawn agents for a HIGH or EPIC complexity task
- A planner has produced a plan that needs validation before execution
- Step 0.6 creation preflight in the router decision workflow
Skip when:
- Task complexity is TRIVIAL or LOW
- Task is a single-file bug fix or documentation update
- User has explicitly bypassed the gate via
READINESS_GATE=off
Readiness Checks
Check 1: Plan Completeness
Every task in the plan must have:
- A clear subject/title
- Acceptance criteria (at least 1 testable condition)
- File paths to create or modify (at least 1)
- Dependency declarations (blockedBy, if any)
Pass: All tasks have all required fields.
Fail: Tasks missing acceptance criteria or file paths.
Check 2: Architecture Compliance
Verify that planned changes align with framework conventions:
- New skills use
.claude/skills/<name>/SKILL.md path pattern
- New agents use
.claude/agents/<category>/<name>.md path pattern
- New hooks use
.claude/hooks/<category>/<name>.cjs path pattern
- New modules follow kebab-case naming
- Gate 4 paths use creator skills (not direct writes)
Pass: All planned paths follow conventions.
Fail: Non-standard paths or direct writes to creator-managed paths.
Check 3: Dependency Graph Validity
- No circular dependencies between tasks
- All blockedBy references point to existing tasks
- Critical path identified (longest chain of dependencies)
Pass: DAG is valid with no cycles.
Fail: Circular dependencies or dangling references.
Check 4: Risk Assessment Present
For HIGH/EPIC tasks:
- Rollback strategy documented
