graph-code

Pass

Audited by Gen Agent Trust Hub on Nov 5, 2026

Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill automates the installation of Python dependencies (tree-sitter, networkx) required for its operation via Node.js scripts. These installations target standard package registries and are triggered only when the graph functionality is enabled in the project configuration.
  • [COMMAND_EXECUTION]: The skill executes various python commands to interact with a local code_graph script for building and querying the database. These commands are utility-focused (e.g., status, build, query, trace) and use strictly defined CLI arguments.
  • [SAFE]: The skill includes extensive documentation and 'blocking' mode dispatch logic to ensure the AI agent follows a structured protocol, requiring verification of graph results through direct file reading.
Audit Metadata
Risk Level
SAFE
Analyzed
Nov 5, 2026, 06:42 PM