Claude Skill MartBrowse skillsQuick linesLearn by videoTerminal guideWhat is a Skill?
 
← Back to list

trace-mcp Code Intelligence Routing

A routing skill that makes Claude use trace-mcp's semantic code tools instead of Read/Grep/Glob, cutting tokens while improving accuracy.

Dev & CodingIntermediate★ 189⑂ 24AI score 8/10Last updated: Oct 11, 2026

What it does

  • Redirects all source-code exploration from Read/Grep/Glob/Bash to trace-mcp's semantic tools (search, get_outline, get_symbol, get_call_graph, and more).
  • Ships a decision matrix mapping each task to the right tool: symbol lookup, change-impact analysis, HTTP request flow, component tree, circular imports, dead code, complexity hotspots.
  • Standardizes the workflow: get_project_map + get_task_context at session start, register_edit after every edit, check_duplication before creating new symbols.
  • Enforces token-efficiency rules: batch independent calls, never re-read a file, read only the target line range before an Edit, no subagent delegation for exploration.
  • Adds secret-safety rails: never cat a .env or dump os.environ/process.env; use get_env_vars, which returns keys and inferred types only.

Who it's for

  • Developers who already run the trace-mcp MCP server with an indexed repository.
  • Teams on large polyglot monorepos where grep-heavy exploration blows up the context window.
  • Leads doing impact analysis, call-graph tracing, or tech-debt audits before refactors.

Example uses

  1. "Fix the refund cancellation logic" → get_task_context pulls exactly the relevant code, then only the target symbol's line range is read.
  2. "What breaks if I change this signature?" → get_change_impact plus find_usages and get_tests_for.
  3. "Which env vars does this project use?" → get_env_vars returns keys and types without ever exposing secret values.

· · · Install guide · · ·

Try it now, no install

Paste this into Claude to use the skill without installing anything.

Read the instructions in this file and follow them to help me:
https://raw.githubusercontent.com/nikolai-vysotskyi/trace-mcp/HEAD/skills/trace-mcp/SKILL.md

What I want: (describe your task here)

If Claude can't open the link, open it yourself and paste the contents instead.

↓ If it works for you, download the ZIP below and install it. Then it runs on its own — no pasting each time.

Install in the Claude app (no terminal)
  1. Download the ZIP with the button below.
  2. In Claude, open Settings → Capabilities and turn on 'Code execution and file creation'. (one time)
  3. Go to Customize → Skills → + → 'Upload a skill' and upload the ZIP.
↓ Download ZIP
Install in Claude Code

Let Claude do it — paste this into Claude Code

Install the skill I found on Claude Skill Mart.
Copy the skills/trace-mcp folder from the GitHub repo nikolai-vysotskyi/trace-mcp into my ~/.claude/skills/trace-mcp/.
When it's done, tell me in one line what this skill can do.

Install with a command instead

git clone https://github.com/nikolai-vysotskyi/trace-mcp.git /tmp/trace-mcp && mkdir -p ~/.claude/skills && cp -r /tmp/trace-mcp/skills/* ~/.claude/skills/

⚠ This is a third-party skill. Check the source repository before installing.

  1. Install the trace-mcp MCP server first and index your repository — the skill does nothing without it.
  2. Clone the repo: git clone https://github.com/nikolai-vysotskyi/trace-mcp.git
  3. Create the skills folder: mkdir -p ~/.claude/skills
  4. Copy the skills in: cp -r trace-mcp/skills/* ~/.claude/skills/ (this also installs trace-mcp-refactoring, trace-mcp-codemod, and trace-mcp-pre-commit).
  5. Restart Claude Code and confirm trace-mcp appears in your skill list.
  6. In a fresh session, ask "give me an overview of this project" — if Claude calls get_project_map, the wiring works.
View source on GitHub ↗License: MIT