codanna

Analyzes code symbols, call graphs, and documentation via Codanna proxy scripts.

Updated Mar 22, 2026
One-click install
npx skills add https://github.com/serguei9090/LogLensAi --skill codanna-serguei9090
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: codanna
Source: https://github.com/serguei9090/LogLensAi/tree/main/.agents/skills/codanna
Command: npx skills add https://github.com/serguei9090/LogLensAi --skill codanna-serguei9090

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes scripts (resource) components.

What problem does it solve? Manually tracing symbol dependencies, callers, and impact across a large codebase is slow and error-prone, and calling raw Codanna MCP commands with hand-written JSON arguments frequently breaks autonomous agent loops due to formatting errors. ## Core Features & Use Cases - Symbol & Dependency Analysis: Find symbols semantically, list what a function calls, find its callers, and map the full impact graph with configurable depth. - Semantic Search: Search code symbols and markdown documentation collections with similarity thresholds and context enrichment. - Safe Proxy Scripts: All operations run through wrapper scripts in scripts/codanna/ that handle JSON formatting and surface Codanna's system_message hints for chaining queries. - Use Case: Before refactoring a core function, run search.py to locate the symbol, callers.py to find dependents, and impact.py to map the blast radius, then re-index with index.py after the change. ## Quick Start Use the codanna skill to find the symbol for the log parsing function and show me its full impact graph.

Frequently Asked Questions about codanna

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I find what functions call a specific symbol?

Run scripts/codanna/callers.py with the symbol name or its symbol_id to list all callers. You can get the symbol_id from a prior search.py result and chain it directly without re-searching.

How to analyze the impact of a code change before refactoring?

Use scripts/codanna/impact.py with the symbol name and an optional --depth argument to generate the full dependency graph. Combine it with calls.py and callers.py to see both directions of the call graph.

Why should I avoid calling codanna mcp commands directly?

Raw MCP calls require hand-written JSON arguments, and formatting errors break the autonomous agent loop. The proxy scripts in scripts/codanna/ handle argument construction and surface Codanna's chaining hints reliably.

How do I re-index the codebase after structural changes?

Run scripts/codanna/index.py with no arguments for a smart full re-index of code and docs. Indexing also runs automatically via lefthook on commit, but manual re-indexing is recommended after refactors.

What similarity threshold should I use for semantic code search?

Results scoring below 0.3 are likely irrelevant. Pass --threshold 0.5 to search.py for cleaner results when searching code symbols or documentation collections.