skills/specificity-judge/SKILL.md
Score whether each task is specific enough for autonomous, one-shot, production-ready execution.
npx skillsauth add bigeasyfreeman/adlc specificity-judgeInstall this skill globally with one command. Works with Claude Code, Cursor, and Windsurf.
3 of 9 scanners reported clean
Some scanners were skipped, did not run, or reported a non-clean status. Review each row below.
Run this at Eval Council Gate 0 after schema validation succeeds.
For each task provide:
artifact_typedecision_contractreference_implfiles_to_modifyfiles_to_createtech_debt_boundariescompatibility_contractevidence_responsibilitiesdefinition_of_donedependenciesverification_spec.primary_verifiertarget_files and expected_failure_mode{
"task_id": "TASK-001",
"score": 0.74,
"rationale": "The task names files and a verifier, but the acceptance criteria still leave rollback evidence implicit.",
"failing_signals": ["missing_compatibility_contract", "ambiguous_user_path"]
}
score >= 0.8 -> pass0.6 <= score < 0.8 -> warnscore < 0.6 -> revise with low_specificitytarget_files, missing expected_failure_mode, and verifier targets that do not pin the intended behavior.decision_contract.status == unresolved or decision_contract.blocks_implementation == true. Those must be decision gates, not executable tasks.scope_lock_epic may lock context, but an implementation task must own the executable work. The same behavior should not be executable in both.tech_debt_boundaries, compatibility_contract, evidence_responsibilities, or definition_of_done; these are one-shot production readiness fields, not optional prose.fast_judge is unavailable for the active runtime, emit stuck with reason specificity_judge_unavailable.unresolved_type1_in_implementationmissing_artifact_typeparent_child_duplicate_scopeunresolved_dependency_aliasmissing_reference_implmissing_target_filesmissing_expected_failure_modemissing_tech_debt_boundariesmissing_compatibility_contractmissing_evidence_responsibilitiesmissing_definition_of_donemissing_validation_taskambiguous_user_pathambiguous_failure_modedevelopment
Orchestration skill: chains the full ADLC Build Loop. PRD → Brief → Council → Scaffold → Codegen → LDD → TDD → Council → PR. Use when implementing a new feature end-to-end.
development
# Skill: Helm & ArgoCD Deployment > Validates Helm charts and generates ArgoCD Application manifests when the ADLC pipeline produces infrastructure or service code. Ensures every deployable artifact has correct chart structure, environment-specific values, and a GitOps-ready Application manifest before code review. --- ## Why This Exists Without deployment validation in the pipeline, common failures slip through to production: - **Helm charts fail `helm template`** because of missing values,
testing
Decide whether an intersecting verifier actually exercises the semantic change.
development
# Skill: UX Flow Builder > Generates user flow diagrams (Mermaid) from PRD personas and screen specifications. Surfaces dead ends, missing screens, and disconnected flows before design or engineering starts. Helps PMs think in screens, not features. --- ## Trigger - Automatically during PRD Phase 4 (Personas & Flows) to visualize the user journey - On-demand when the PM says "show me the flow" or "map the user journey" - During PRD evaluation to verify screen connectivity --- ## Input ```