The Agent Skills Directory

[COMMAND_EXECUTION]: The skill performs shell command execution to facilitate its analysis and verification processes. This includes repository enumeration using git ls-files and find, as well as the execution of testing and linting tools defined within the project, such as pytest, cargo test, and scripts found in package.json.
[INDIRECT_PROMPT_INJECTION]: The skill processes all files within the repository to identify code decay and technical debt, creating a broad attack surface for indirect prompt injection if the analyzed codebase contains malicious instructions.
Ingestion points: The skill reads all files within the user-defined or auto-detected scope during its four analysis dimensions (SKILL.md; sweep-guide.md Step 2).
Boundary markers: The instructions do not specify the use of delimiters or 'ignore' instructions when reading file content for analysis, which may lead the agent to follow instructions embedded in the code.
Capability inventory: The skill is capable of modifying files (Edit/Write) and executing arbitrary commands derived from project configurations (sweep-guide.md Step 2d, 2e).
Sanitization: There is no evidence of sanitization, filtering, or validation of the content ingested from the codebase before it is used to generate fixes.

brooks-sweep