菜单
Pre-completion protocol for implementation agents — gather context, dispatch the pre-completion-reviewer subagent, and handle its report before writing stage notes and recommending /ship-issue. Load at the end of /tdd-plan and /build-plan after all implementation steps are complete.
用 Codex 或 Claude 帮你安装 复制这段 Prompt，粘贴到 Codex、Claude 或其他助手里，让它检查 Skill 页面并帮你完成安装。
基于 SOC 职业分类
Anthropic Claude Pro/Max OAuth compatibility workflow for this repo. Use when debugging Anthropic OAuth failures, misleading extra-usage errors, Pi request shaping, prompt fingerprinting, or deciding between hook-based fixes and deeper provider overrides.
Codebase intelligence via fallow CLI — dead code, duplication, complexity, refactoring targets. Load when investigating unused code, planning refactors, or reviewing fallow output.
Heuristics and process for discovering structural improvements in this package. Load when planning a new improvement round — contains the smell taxonomy, analysis workflow, and prioritization framework distilled from many phases of refactoring.
TypeScript conventions, code design principles (SOLID, self-documenting code, file organization), structural design heuristics (dependency width, LoD, output arguments), pnpm rules, ES2024 target, Pi SDK patterns, and Biome/ESLint conflict workarounds. Load during implementation, refactoring, or code review.
Project-specific markdown rules (one-sentence-per-line, compact tables, sequential numbering) and YAML frontmatter schema for plans/retros. Load when writing or editing markdown — contains rules that differ from standard markdownlint defaults.
Vitest mock patterns (vi.mock, vi.hoisted, vi.fn reset), TDD planning rules, and general test strategy. Load when writing or debugging tests.
| name | pre-completion |
| description | Pre-completion protocol for implementation agents — gather context, dispatch the pre-completion-reviewer subagent, and handle its report before writing stage notes and recommending /ship-issue. Load at the end of /tdd-plan and /build-plan after all implementation steps are complete. |
Load this skill when all implementation steps are complete and the final deterministic checks have passed. Follow these steps in order before moving to "Summarize" and "Write stage notes."
Run from the repo root:
BASE=$(git describe --tags --abbrev=0 2>/dev/null || git rev-list --max-parents=0 HEAD)
git diff --name-only $BASE..HEAD
Note:
issue: field or the plan filename pattern NNNN-).docs/plans/NNNN-*.md matching the issue number; may be absent for unplanned work).Dispatch the pre-completion-reviewer subagent via the subagent tool:
subagent_type: "pre-completion-reviewer"description: "Pre-completion review for issue #N"prompt: include the issue number, the modified-files list from Step 1, and the plan file path.Example prompt to pass:
Review issue #236.
Plan file: docs/plans/0236-pre-completion-reviewer.md
Modified files since last tag:
.pi/agents/pre-completion-reviewer.md
.pi/skills/pre-completion/SKILL.md
.pi/prompts/tdd-plan.md
src/oauth-transport.ts
AGENTS.md
Wait for the reviewer to complete and return its report before continuing.
If you cannot read an explicit Overall: PASS|WARN|FAIL line, treat it as "report not captured" and re-dispatch per Step 2.
Do not proceed to "Summarize" on an uncaptured or banner-only result.
Proceed to the "Summarize" step in the template. Include the one-line verdict in the stage notes ("Pre-completion reviewer: PASS").
Proceed to "Summarize."
Include the verdict and WARN findings in the stage notes under a "Reviewer warnings" line.
The user can decide whether to address warnings before running /ship-issue.
Stop — do not proceed to "Summarize" or "Write stage notes." Report the reviewer's "Fix required" block to the user and ask how to proceed:
Do not proceed automatically — let the user decide.