The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted project files which may contain embedded instructions.
Ingestion points: Files are read via Read, Grep, and Glob tools across the project directory (SKILL.md, references/ruff-rules-reference.md).
Boundary markers: None identified. The skill directly reads code and configuration files into the agent's context.
Capability inventory: Uses Bash to run linters and Write/Edit to apply auto-fixes to the codebase.
Sanitization: No explicit sanitization or instruction to ignore comments/text within the analyzed code files is provided.
[DYNAMIC_EXECUTION]: The reference files contain pre-defined Python one-liners used to process JSON output from tools.
Evidence: references/biome-rules-reference.md and references/ruff-rules-reference.md include commands like python3 -c "import json,sys; ..." to filter linter diagnostics.
Risk: While the scripts are static and utility-oriented, they represent the generation and execution of code at runtime.
[COMMAND_EXECUTION]: The skill makes extensive use of the Bash tool to execute shell commands.
Evidence: Instructions include running ruff, npx @biomejs/biome, and git diff via the shell.
Context: These are standard operations for a linting and formatting skill.

code-linting