AGENT SKILL

install-and-setup

Install, configure, update, serve, and troubleshoot code-review-graph CLI/MCP graph setup across supported coding-agent platforms.

License 许可
MIT
最近更新
2026-09-03
Stars
★ 351
Forks
27

01 前言 Front Matter

name
install-and-setup
license
MIT
metadata
{ "disco-role": "operating" }
description
Install, configure, update, serve, and troubleshoot code-review-graph CLI/MCP graph setup across supported coding-agent platforms.
disable-model-invocation
true

02 资源文件 3

03 SKILL.md 正文

Install and Setup

Use this sub-skill when the task is to install code-review-graph, create or refresh its local graph database, configure MCP clients or hooks, start the server, inspect status, visualize the graph, or diagnose setup failures.

Start here

  1. Confirm Python 3.10+ and install the package:
    pip install code-review-graph
    # or: pipx install code-review-graph
    # or: uvx code-review-graph --help
    
  2. Configure an AI coding platform from the target repository root:
    code-review-graph install
    
  3. Build the graph once:
    code-review-graph build
    
  4. Verify:
    code-review-graph status
    
  5. If a future session cannot see the MCP server, restart the client/editor and verify the session opened in the same repository root.

Read references/workflows.md for platform setup, graph lifecycle, server, visualization, uninstall, and environment-check workflows. Read references/troubleshooting.md when a command is missing, a graph is stale, MCP does not start, HTTP serving is rejected, or Python/package resolution is unclear.

Route by task

User task Do this
"Install CRG in this repo" Run code-review-graph install, then code-review-graph build, then code-review-graph status.
"Only configure Codex/Cursor/Claude/etc." Use code-review-graph install --platform <platform> with a supported platform name.
"The graph is stale" Run code-review-graph update --brief; use build for a full rebuild after major branch switches or parser/schema concerns.
"Run the MCP server" Use code-review-graph serve for stdio or code-review-graph serve --http --host 127.0.0.1 --port 5555 for loopback HTTP.
"Watch the repo" Use code-review-graph watch for one repo or route multi-repo daemon tasks to integrations-and-extensions.
"Show graph stats" Use code-review-graph status or code-review-graph status --json.
"Visualize the graph" Use code-review-graph visualize after a graph exists; open the generated local HTML.
"Remove CRG" Use code-review-graph uninstall --dry-run first; add --yes only after confirming the preview.

Supported setup surfaces

install/init can configure supported platforms including Codex, Claude Code, CodeBuddy Code, Cursor, Windsurf, Zed, Continue, OpenCode, Antigravity, Gemini CLI, Qwen, Kiro, Qoder, GitHub Copilot, and GitHub Copilot CLI. Platform config paths and generated hook/skill behavior are summarized in references/workflows.md.

The graph database is project-scoped by default under .code-review-graph/graph.db. The Python package is user-scoped, while MCP config and hooks are normally project- or client-scoped depending on platform.

Safe bundled helpers

Boundaries

  • For reviewing changed code, use review-changes.
  • For structural exploration, search, architecture, or refactor previews, use graph-exploration.
  • For embeddings, custom languages, GitHub Action setup, multi-repo registry, daemon, wiki, or eval workflows, use integrations-and-extensions.

Verification anchors

Native tests that ground this sub-skill include install lifecycle tests, CLI build/update/status tests, MCP server wiring tests, HTTP Host/Origin guard tests, and status/stat regressions. Run selected native tests only after the whole repo skill is integrated; do not use them as runtime dependencies.

04 同仓库更多 Skills

VectorSpaceLab
VectorSpaceLab/AREX-Skill

passnet-triton-opt

Make a single PassNet Triton kernel fast AND numerically faithful: performance model (when a replacement can win at all), block/grid/warp tuning, autotune policy, launch overhead, and per-op numeric recipes to pass the dtype baseline tolerances. Use when a pass matches and is correct but speedup below expected, or when correctness fails by small numeric margins.

★ 351
VectorSpaceLab
VectorSpaceLab/AREX-Skill

passnet-solve

END-TO-END playbook for solving one PassNet sample: analyze the computation graphs, decide the optimization strategy (what to fuse, what to replace, what to leave alone), drive the iteration loop, and maximize the sample score. This is the ENTRY skill — invoke it first for any PassNet optimization task; it tells you when to use passnet-pattern-fusion, passnet-triton-opt and passnet-feedback.

★ 351
VectorSpaceLab
VectorSpaceLab/AREX-Skill

passnet-skill

PassNet GPU kernel optimization via compiler passes. Design and implement Triton-based optimization passes, create pass files under ./pass_dir/, self-evaluate with pass_evaluator, and iterate to maximize GPU speedup.

★ 351
VectorSpaceLab
VectorSpaceLab/AREX-Skill

passnet-pattern-fusion

Author PassNet pass files: write patterns that actually MATCH the FX graph, pick fusion regions (which ops to absorb into one Triton kernel), and structure multi-pass submissions with the mandatory shared-dispatch architecture. Use when creating or fixing pattern()/replacement_args()/replacement_func() files, when a pass "failed to match", or when deciding how to fuse multiple kernels/ops.

★ 351
VectorSpaceLab
VectorSpaceLab/AREX-Skill

passnet-orchestrate

Round/batch planning and decision support for PassNet graph optimization. Use for multi-sample triage, worker allocation, eval-budget policy, and post-eval keep/revert decisions. For solving one concrete sample end to end, use passnet-solve as the entry skill; consult this skill only when you need broader planning or a specific decision gate.

★ 351
VectorSpaceLab
VectorSpaceLab/AREX-Skill

passnet-feedback

Fast, deterministic feedback for PassNet work: pre-flight pattern/match verification WITHOUT burning a GPU evaluation, per-node bottleneck analysis of a sample's graphs, and eval-log parsing into per-variant status + estimated score + failure classification. Use BEFORE every GPU evaluation (check_pattern), at the START of a sample (analyze_graph), and AFTER every evaluation (parse_eval_log).

★ 351