← ClaudeAtlas

handoff-summarylisted

Emit the Build Handoff Summary artifact in its fixed format. Use at the end of a build, immediately after /3p-review passes — invoked by the build-model workflow (dedicated build model) or by the main model after all build phases. Produces the fixed handoff template the reviewing model consumes.
nikhilw/structured-agentic-workflow · ★ 3 · AI & Automation · score 79
Install: claude install-skill nikhilw/structured-agentic-workflow
# Handoff Summary You are emitting the **Build Handoff Summary** — the single artifact that carries a completed, reviewed build to the next reader. > **Output style:** Check memory for `workflow-config:caveman-level`. If set, adapt your output brevity to that level while preserving technical accuracy. ## When this runs This skill is invoked **after** the build is complete and **after** `/3p-review` has passed. It does not build, test, or review — it only produces the handoff record. If `/3p-review` has not run yet, stop and run it first. `/3p-review` is its own artifact — the handoff summary does not restate or reference its findings. ## Your only job Follow the template below: same headings, same order, same casing. Fill each section from the build record. Do **not** rename sections, add sections, drop sections, or replace the template with your own prose. This is a fixed artifact format so the consuming model can parse it reliably. ```markdown ## Build Handoff Summary **Plan:** [plan file path] ### Deviations - **Phase N: [Name]** — [what changed and why, one line] - (Only list phases that deviated from the plan. If nothing deviated, write "None.") ### Verification Runs - **[plan criterion, or a plain name for the run]** → exit [code] — [N passed, M failed, K skipped] - (Name the run; do not write out the shell line. The full-suite run is mandatory here. Report the counts the terminal printed, not what you expected.) ### Unproven Criteria - **[plan criterion]** —