End-of-session knowledge cleanup with OCD-level rigor — reconciles project docs (CLAUDE.md, README.md, docs/) and agent memory against the code so nothing rots. 会话结束后对项目文档和记忆进行洁癖级审查与同步。MUST trigger when the user says: "sync up", "tidy up docs", "update memory", "clean up docs", "/sync", "/neat", "同步一下", "整理文档", "整理一下", "更新记忆", "梳理一下", "收尾", "这个阶段做完了", "新人能直接上手", or any phrase suggesting a dev milestone where knowledge needs reconciliation. Also trigger when the user reports stale docs, conflicting memories, or wants a clean handoff to teammates or other agents. Bare "整理" / "tidy" with prior dev context counts — do not under-trigger. Cross-platform: works on Claude Code, OpenAI Codex, OpenCode, and OpenClaw.
90
90%
Does it follow best practices?
Impact
Pending
No eval scenarios have been run
Passed
No known issues
Quality
Discovery
89%Based on the skill's description, can an agent find and select it at the right time? Clear, specific descriptions lead to better discovery.
This is a strong skill description with excellent trigger term coverage (bilingual, slash commands, contextual triggers) and clear completeness. The main weakness is that the core capability description could be more specific about the concrete actions performed during reconciliation. The colorful language ('OCD-level rigor', '洁癖级审查') adds personality but slightly detracts from precision, and the cross-platform mention at the end is somewhat tangential to skill selection.
Suggestions
Replace 'reconciles project docs and agent memory against the code' with more specific actions, e.g., 'removes stale entries from CLAUDE.md, updates README.md to reflect current architecture, prunes outdated agent memories, flags doc-code inconsistencies'.
| Dimension | Reasoning | Score |
|---|---|---|
Specificity | The description names the domain (project docs, agent memory) and the core action (reconciles project docs against code), but the specific actions are somewhat vague — 'reconciles' and 'cleanup' are high-level. It mentions specific files (CLAUDE.md, README.md, docs/) which adds concreteness, but doesn't list granular actions like 'removes stale entries', 'updates outdated references', etc. | 2 / 3 |
Completeness | Clearly answers both 'what' (reconciles project docs and agent memory against code to prevent rot) and 'when' (explicit 'MUST trigger when' clause with extensive trigger phrases and contextual conditions like stale docs, milestone completion, and handoff scenarios). | 3 / 3 |
Trigger Term Quality | Excellent coverage of natural trigger terms in both English and Chinese, including slash commands (/sync, /neat), natural phrases ('sync up', 'tidy up docs', 'update memory', 'clean handoff'), and Chinese equivalents ('同步一下', '整理文档', '收尾'). Also covers contextual triggers like 'stale docs' and 'conflicting memories'. Very thorough. | 3 / 3 |
Distinctiveness Conflict Risk | Highly distinctive niche — end-of-session knowledge cleanup specifically targeting project documentation and agent memory reconciliation against code. The specific file targets (CLAUDE.md, README.md, docs/), the session-end context, and the extensive unique trigger terms make it very unlikely to conflict with other skills. | 3 / 3 |
Total | 11 / 12 Passed |
Implementation
85%Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.
This is a well-structured, highly actionable skill for knowledge base synchronization. Its greatest strengths are the concrete five-step workflow with mandatory validation checklist, the clear three-audience mental model, and appropriate progressive disclosure to reference files. The main weakness is some verbosity in the motivational sections and concept explanations that Claude doesn't need, though the operational content is tight and executable.
Suggestions
Trim or remove the '为什么这件事重要' section — Claude doesn't need motivation about why stale docs are bad; the three-audience table alone conveys the key insight.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The skill contains some unnecessary motivational/philosophical content ('为什么这件事重要' section explains things Claude already understands about stale docs). The three-audience table is valuable but the surrounding explanation is verbose. However, the operational sections are reasonably tight. | 2 / 3 |
Actionability | Provides concrete, executable commands (ls, find, grep), specific file paths, exact editing principles, and a detailed checklist. The workflow steps include real shell commands and specific file names rather than vague descriptions. The change summary template is copy-paste ready. | 3 / 3 |
Workflow Clarity | Five clearly sequenced steps with explicit validation (Step 4 is a mandatory self-check checklist with specific grep commands). The workflow includes feedback loops ('哪条打不了勾,回去补'), a priority ordering for edits (docs first, then project root, then memory), and handles error recovery for contradictions. The 'change impact matrix' thinking in Step 2 is an excellent checkpoint. | 3 / 3 |
Progressive Disclosure | Core workflow is self-contained in the SKILL.md with clear references to two supplementary files (references/sync-matrix.md and references/agent-paths.md) that are one level deep and well-signaled. The inline content covers the common cases while deferring exhaustive mappings and platform-specific paths to reference docs. | 3 / 3 |
Total | 11 / 12 Passed |
Validation
100%Checks the skill against the spec for correct structure and formatting. All validation checks must pass before discovery and implementation can be scored.
Validation — 11 / 11 Passed
Validation for skill structure
No warnings or errors.
Reviewed
Table of Contents