Find Similar Code

Find Similar Code

Uses semantic vector search for intelligent matching across the codebase.

Inputs

Workflow

1. Check Known Issues

• check_known_issues(tool_name="code_search", error_text="")

2. Detect Project

• Infer from cwd; default automation-analytics-backend

3. Semantic Search

• code_search(query=query, project=project, limit=limit)
• Parse results: file, line, snippet

4. Get Related Patterns

• knowledge_query(project=project, section="patterns.coding")
• Include top 3–5 in output

5. Build Summary

Output markdown with:

• Query and project
• Found N Results: each with file:line, snippet
• Related Coding Patterns: from knowledge
• Tip: Use code_search() for more control

6. Error Handling

• If "index not found": learn_tool_fix("code_search", "index not found", "Vector index not created", "Run skill_run('bootstrap_knowledge')")
• If "empty index": suggest skill_run('knowledge_refresh', '{"full_rescan": true}')

7. Log

• memory_session_log("Code search in {project}", "Query: {query[:50]}, Results: {count}")

Key MCP Tools

• code_search — semantic vector search
• knowledge_query — related patterns
• check_known_issues, learn_tool_fix — error handling
• memory_session_log — session logging