The Agent Skills Directory

[PROMPT_INJECTION]: The skill processes untrusted codebases through indexing and file reading, creating a surface for indirect prompt injection where malicious instructions embedded in the analyzed code could potentially influence the agent's behavior. * Ingestion points: Content is read from the user-provided project_path via the mcp__hex-graph__index_project and mcp__hex-line__read_file tools. * Boundary markers: There are no explicit delimiters or instructions provided to the agent to disregard instructions found within the project files. * Capability inventory: The skill uses tools for structural code analysis and file reading, though it does not include direct execution capabilities for the analyzed code. * Sanitization: No explicit sanitization or validation of the processed code content is described in the skill workflow.

ln-020-codegraph