The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted codebase files that could contain hidden instructions designed to override agent behavior during the audit process.
Ingestion points: Codebase files accessed via Read, Grep, and Glob tools.
Boundary markers: No explicit delimiters or instructions to ignore embedded content within the analyzed files are provided.
Capability inventory: The skill uses Bash to run linters and coverage tools, and Read for file content analysis.
Sanitization: The skill does not mention sanitizing or escaping the content of the files before processing them.
[COMMAND_EXECUTION]: The skill uses the Bash tool to execute external diagnostic utilities such as linters and test coverage tools. While this is expected functionality for a code auditor, it provides a command execution surface that could be targeted if the tool arguments are influenced by file content.

code-auditor