The Agent Skills Directory

[COMMAND_EXECUTION]: The script detect-and-fix.sh executes a range of command-line tools based on the project's structure. It uses the eval command to run these tools from predefined strings, which is a sensitive execution pattern.
[EXTERNAL_DOWNLOADS]: The skill uses npx to execute Node.js tools such as @biomejs/biome, eslint, and prettier. This can result in downloading packages from the npm registry if they are not already installed locally.
[REMOTE_CODE_EXECUTION]: By using npx and eval, the skill has a surface for executing code that originates from external registries or is triggered by project-specific configuration files.
[COMMAND_EXECUTION]: The skill possesses an indirect prompt injection surface as it processes untrusted project code and configuration files using powerful command-line utilities.
Ingestion points: Reads project configuration files (e.g., biome.json, .eslintrc, Cargo.toml) and source code files in the detect-and-fix.sh script.
Boundary markers: None present.
Capability inventory: Execution of arbitrary developer tools including npx, cargo, and go via the eval command in a bash script.
Sanitization: No explicit sanitization or validation of project files or tool outputs before execution or reporting.

code-lint-fix