jcodemunch-mcp-code-retrieval

Pass

Audited by Gen Agent Trust Hub on May 16, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill requires the installation of the jcodemunch-mcp package from PyPI via pip or uvx. It also suggests optional dependencies such as sentence-transformers and torch for semantic search functionality.\n- [COMMAND_EXECUTION]: Installation and configuration steps involve executing shell commands like uvx jcodemunch-mcp and the jcm CLI helper to integrate the tool with various IDEs and agents.\n- [PROMPT_INJECTION]: The skill has an attack surface for indirect prompt injection as it processes content from local repositories that may contain adversarial instructions.\n
  • Ingestion points: Local source code files indexed via the index_repository tool in SKILL.md.\n
  • Boundary markers: The skill does not explicitly define delimiters or warnings to ignore embedded instructions when retrieving code snippets.\n
  • Capability inventory: The skill provides access to code implementations and class hierarchies but does not execute the indexed code or perform network operations with the retrieved data.\n
  • Sanitization: Code is parsed using tree-sitter AST, which provides a more robust structural representation than raw text, though it does not verify the intent of the retrieved code comments or documentation.
Audit Metadata
Risk Level
SAFE
Analyzed
May 16, 2026, 08:19 PM
Security Audit — agent-trust-hub — jcodemunch-mcp-code-retrieval