sanyuan-skills
Six Claude Code skills: code review for SOLID and security, the Sigma tutor, skill audit and creation, a wiki from notes and a book-reading coach
- 3,966 ★ on the repository
- Last commit
- MIT
Extended access
Why this level
- The review skill runs git and repository search and reads all of your changes
- The skill-forge skill contains Python scripts that create and package skills and write files to disk
Install
- Works in
In your terminal, with SkillFoxx CLI
npx skillfoxx add skills/sanyuan-skillsDetects the agents on your machine, shows the access level and pins the version.
Other ways to install
Access and checks
How we reviewExtended accessThe tool runs code or makes network calls and only reads external data.
- The review skill runs git and repository search and reads all of your changes
- The skill-forge skill contains Python scripts that create and package skills and write files to disk
This is third-party code. Review the repository files before installing.
What it does
sanyuan-skills is a set of six skills by sanyuan0704, installed one at a time with npx skills add. The main skill, code-review-expert, reviews your current git changes through a senior engineer's eyes. It looks at SOLID violations, removal candidates, security gaps (XSS, injection, SSRF, missing authorization checks, races) and code quality: error handling, performance and boundary conditions. Findings get levels from P0 (blocks the merge) to P3 (optional), and by default the skill only writes a report and changes nothing until you ask. The other skills: Sigma tutors with a Bloom mastery-learning scheme and Socratic questions, skill-review audits the structure and description of other skills, skill-forge creates new ones using 12 techniques, wiki-ingest compiles articles and notes into a cross-referenced wiki, and book-study works as a reading coach with spaced repetition.
Who it is for. For developers who want to check their own changes before a peer review, and for anyone who writes and refines agent skills.
Good fit when
- Before opening a pull request you want the diff checked for SOLID, security and error handling
- A large diff of over 500 lines needs to be reviewed file by file and by module rather than in one stream
- You want a removal plan for dead code marked "now" or "defer"
- You wrote a skill and want its structure, description and token use checked
- You need a tutor that asks questions about a topic instead of giving ready answers
Not a fit when
- You need automatic comments directly in a GitHub pull request: the skill reads a local git diff and writes a report
- You need a static analyzer in CI: the agent does the review, the result depends on the model and does not replace linters
- The repository has no changes: on an empty git diff the skill asks whether to review staged changes or a commit range
- You only want one set of code rules installed: here are six different skills
Example request
/code-review-expert Review the current changes, especially security and error handling, and change nothingLimitations
The review skill reads git status and git diff and, when needed, searches the repository with rg or grep, so quality depends on the agent model and the size of the diff. The author built the set for Claude Code and "other agent terminals"; the description lists no verified environments. Installation uses the npx skills add command from an external package, so Node.js is needed. Some skills, such as Sigma and book-study, are designed for a dialogue rather than a one-off request.
How to disable. Delete the folder of the skill you want to remove from the agent skills directory, for example skills/code-review-expert. The other skills in the set are independent and keep working.
FAQ
How do I install code-review-expert in Claude Code?
Run npx skills add sanyuan0704/sanyuan-skills --path skills/code-review-expert. Then call /code-review-expert in a session with uncommitted changes. The other skills in the set install the same way with a different --path value.
What do the P0–P3 levels in the report mean?
P0 is a critical problem such as a vulnerability or data loss and blocks the merge. P1 is a logic error or a significant SOLID violation. P2 is a code smell to fix now or file as a follow-up. P3 covers style and naming and is optional.
Does the skill fix the code itself?
No by default: it works in report mode. It makes changes only after you ask.
Which security checks does the review include?
The skill looks for XSS, SQL and command injection, SSRF, path traversal, authorization gaps, secrets leaking into logs and files, unbounded loops, unsafe deserialization, weak crypto and races under concurrent access. For each finding it describes exploitability and impact.
What does skill-forge do?
It is a meta-skill for creating skills with 12 proven techniques. It holds guides on architecture, descriptions and output patterns plus three Python scripts: init_skill.py, package_skill.py and quick_validate.py.
Discussion
No questions or reviews yet.
Related
SkillsClaude Code Skills, Chinese collection
Claude Code Skills 中文精选集
A set of 21 original Claude Code skills with Chinese output: code review, README, API tests, refactoring, profiling
14 skills that find the gaps in a task before the code: interviews, plans, assumption tests and proof that a bug fix works
A set of Claude Code skills for the dev workflow: worktree isolation, a service layer, evidence and review
A code review skill across many languages and frameworks: architecture, performance, security, anti-patterns and common bugs
110 skills for Claude Code and Codex and seven Claude Code agents: development, product, UI, diagnostics and operations, installed whole or as a bundle
Guard skills for coding agents: a second pass over the diff that catches typical AI mistakes in code, tests and docs
Comparison with similar tools
Entries that share the main search query, and how they differ.
| Entry | Official | Authentication | Repository stars | Main difference |
|---|---|---|---|---|
| sanyuan-skills (this entry) | no | 3,966 | ||
| brooks-lint | no | 1,550 | a review plugin grounded in classic engineering books with source references | |
| code-review-graph | no | not required | 32,463 | an MCP server that builds a code graph and gives the agent a map of the repository |
| ielts-claude-skills | no | 330 | four free Claude Code skills |
README in short
The description briefly presents the set as production-grade skills for Claude Code and other agent terminals. A table lists the six skills with a description and the install command through npx skills add with a --path parameter. The quick start shows the invocation commands: /code-review-expert, /sigma, /skill-review, /skill-forge, /wiki-ingest and /book-study.
README badge
Original SKILL.md
--- name: code-review-expert description: "Expert code review of current git changes with a senior engineer lens. Detects SOLID violations, security risks, and proposes actionable improvements." --- # Code Review Expert ## Overview Perform a structured review of the current git changes with focus on SOLID, architecture, removal candidates, and security risks. Default to review-only output unless the user asks to implement changes. ## Severity Levels | Level | Name | Description | Action | |-------|------|-------------|--------| | **P0** | Critical | Security vulnerability, data loss risk, correctness bug | Must block merge | | **P1** | High | Logic error, significant SOLID violation, performance regression | Should fix before merge | | **P2** | Medium | Code smell, maintainability concern, minor SOLID violation | Fix in this PR or create follow-up | | **P3** | Low | Style, naming, minor suggestion | Optional improvement | ## Workflow ### 1) Preflight context - Use `git status -sb`, `git diff --stat`, and `git diff` to scope changes. - If needed, use `rg` or `grep` to find related modules, usages, and contracts. - Identify entry points, ownership boundaries, and critical paths (auth, payments, data writes, network). **Edge cases:** - **No changes**: If `git diff` is empty, inform user and ask if they want to review staged changes or a specific commit range. - **Large diff (>500 lines)**: Summarize by file first, then review in batches by module/feature area. - **Mixed concerns**: Group findings by logical feature, not just file order. ### 2) SOLID + architecture smells - Load `references/solid-checklist.md` for specific prompts. - Look for: - **SRP**: Overloaded modules with unrelated responsibilities. - **OCP**: Frequent edits to add behavior instead of extension points. - **LSP**: Subclasses that break expectations or require type checks. - **ISP**: Wide interfaces with unused methods. - **DIP**: High-level logic tied to low-level implementations. - When you propose a refactor, explain *why* it improves cohesion/coupling and outline a minimal, safe split. - If refactor is non-trivial, propose an incremental plan instead of a large rewrite. ### 3) Removal candidates + iteration plan - Load `references/removal-plan.md` for template. - Identify code that is unused, redundant, or feature-flagged off. - Distinguish **safe delete now** vs **defer with plan**.