一键导入
zeroclaw-simplify-review
ZeroClaw-specific simplify preflight. Checks reuse, known project pitfalls, and avoidable complexity before commit or maintainer review.
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。
菜单
ZeroClaw-specific simplify preflight. Checks reuse, known project pitfalls, and avoidable complexity before commit or maintainer review.
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
Write and execute PostgreSQL or Supabase SQL queries from natural-language requests. Use when the user wants to inspect schema, run read queries, validate data, or safely perform confirmed database changes.
Use Xquik for X/Twitter search, profile tweets, followers, account monitors, webhook events, MCP access, SDK guidance, and extraction workflows through public docs and a user-configured API key.
Automated API testing with schema validation and load testing. Test REST and GraphQL APIs by sending requests, validating responses against schemas, checking edge cases, and reporting failures clearly. Use when the user wants to test APIs, validate endpoints, run load tests, or check schema compliance.
Autonomous code generation agent. Reads context, writes code, runs tests. Takes a task description and produces working, production-quality code changes. Use when the user wants to implement features, write code, or make code changes autonomously.
AI-powered code review. Catches bugs, suggests improvements, checks style. Analyzes code changes, diffs, or pull requests and provides actionable, prioritized feedback. Use when the user wants a code review, wants to check code quality, or needs feedback on a diff or PR.
Analyze CSV, JSON, and SQL data. Generate charts and insights locally. Loads, explores, transforms, and visualizes structured data to deliver clear, actionable insights. Use when the user wants data analysis, visualization, statistical summaries, or anomaly detection.
| name | zeroclaw-simplify-review |
| description | ZeroClaw-specific simplify preflight. Checks reuse, known project pitfalls, and avoidable complexity before commit or maintainer review. |
| version | 0.1.0 |
| author | zeroclaw-labs |
| license | MIT |
| category | coding |
| tags | ["Official"] |
| permissions | ["file_read"] |
Use this skill as a ZeroClaw-specific simplify preflight. It is not a replacement for a general code review skill. Use it before commit, before PR creation, during maintainer review, or when the user asks whether a ZeroClaw change can be simpler, better aligned with existing architecture, or safer against known project pitfalls.
For broad, repository-agnostic code review, use a general code review skill. This skill answers narrower ZeroClaw questions:
The default mode is read-only. Do not edit files, post comments, add labels, submit reviews, or otherwise mutate project state unless the user explicitly asks for that action and confirms the specific scope.
Identify the review target from the user's request, supplied diff, PR context, changed-file list, or specific file paths. Then read the ZeroClaw project instructions that are present in the working tree, especially:
AGENTS.mdCONTRIBUTING.md.github/pull_request_template.mddocs/book/src/contributing/pr-review-protocol.mddocs/book/src/maintainers/reviewer-playbook.mdOnly load the docs that exist and are relevant to the target. If the target is small, prefer the nearest source files, tests, and trait/factory wiring over broad documentation reads.
Run these three passes. If the runtime supports parallel subagents and the user has approved delegation, run them in parallel. Otherwise, run the same passes sequentially.
Check whether the change duplicates existing behavior, bypasses traits or factory wiring, adds speculative abstractions, misses an existing helper, or violates ZeroClaw's trait-driven architecture and crate stability tiers.
Check for bugs, edge cases, stale names, call-site mismatches, state divergence, empty-versus-absent value mistakes, persistence round trips, trait default traps, localization requirements, and missing behavior-first tests.
Check whether the code is unnecessarily complex, allocates or clones more than needed, adds heavy dependencies, expands runtime footprint, hides side effects, or can be made clearer without changing behavior.
Merge and de-duplicate the pass results. Lead with findings, ordered by severity:
Each finding should include a file and line reference when possible, the reason it matters, and a concrete fix direction. If no issues are found, say that clearly and list any remaining test or validation gaps.