The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses common command-line utilities to inspect the repository state and file contents.
Evidence: The file references/inspection-commands.md provides templates for rg (ripgrep), wc (word count), and git diff commands.
Context: These tools are used locally to gather information for the audit process.
[PROMPT_INJECTION]: The skill possesses a surface for indirect prompt injection due to its core function of processing untrusted data (other skill instructions).
Ingestion points: The workflow reads target SKILL.md files, AGENTS.md files, and plugin manifests from the repository.
Boundary markers: The instructions do not define specific delimiters or warnings to ignore embedded instructions within the files being reviewed.
Capability inventory: The skill utilizes rg and git for search/comparison and explicitly includes a step to 'patch' or edit files (Step 5 in SKILL.md).
Sanitization: No explicit content sanitization or validation of external text is mentioned in the provided scripts or instructions.

skill-review