← ClaudeAtlas

implement-epiclisted

Implement one complete PRD epic (`EP-NNN`) through dependency-ordered story slices, one consolidated validation pass, risk-triggered review, and status roll-up. Use when asked to implement an epic, execute an EP-NNN from a PRD, or progress a PRD epic end to end without story-by-story session overhead.
arthjean/skills · ★ 3 · AI & Automation · score 66
Install: claude install-skill arthjean/skills
# implement-epic Implement: $ARGUMENTS ## Objective Complete one epic with the smallest evidence, review, and verification loop that can prove it works. `SCOPE -> IMPLEMENT -> REVIEW -> VERIFY -> STATUS` Print one compact progress line per phase. Keep a `Proof Ledger` under 250 tokens containing: epic ID, ordered incomplete stories, acceptance proof, changed files, risk flags, commands run, status path, and unresolved blockers. Update it instead of repeating plans. ## Profiles Select the lowest profile that covers the highest-risk incomplete story. An explicit profile may escalate but never hide a DEEP condition. | Profile | Use when | Review | Verification budget | |---|---|---|---| | FAST | Clear in-repo pattern, isolated behavior, no sensitive boundary | Diff self-check only | At most 2 commands | | DEFAULT | Shared integration, public behavior, multi-module change, or new dependency using familiar patterns | One orchestrator review | At most 4 commands | | DEEP | Auth, payments, PII, destructive data operations, migration, untrusted shell/file/network input, crypto, LLM tools, cross-service behavior, or genuinely unfamiliar logic | One orchestrator review plus at most one independent review axis; add a second axis only for a distinct security surface | At most 6 commands | Count an aggregate project script as one command even when it runs several tools. File count alone does not select DEEP. ## Phase 1 - SCOPE 1. Parse the PRD path, epic ID, and optional `--pro