The Agent Skills Directory

[PROMPT_INJECTION]: The skill presents an Indirect Prompt Injection surface (Category 8) because it ingests and processes untrusted data from the codebase being audited.
Ingestion points: The skill reads source code and configuration files using Read, Grep, and specialized tools like mcp__hex-line__read_file to perform its analysis.
Boundary markers: File content is analyzed by the agent without explicit boundary markers or instructions to ignore embedded directives in the audited content.
Capability inventory: The skill has access to powerful auditing tools including Bash, Grep, and Glob, which could be targeted if the agent were manipulated by malicious comments or strings within the audited code.
Sanitization: Content from the audited codebase is not sanitized before being evaluated by the agent's logic.

ln-623-code-principles-auditor