The Agent Skills Directory

[PROMPT_INJECTION]: The skill processes untrusted project documentation and source code to generate reports, creating a surface for indirect prompt injection.
Ingestion points: Markdown files and project source code accessed via Glob, Grep, and Read tools.
Boundary markers: No explicit delimiters or instructions to ignore embedded commands were found in the processing logic.
Capability inventory: The skill can execute shell scripts (Bash), write files (Write), and perform codebase searches (Grep).
Sanitization: The tool extracts specific patterns (links, versions) but does not explicitly sanitize the text of the findings before presenting them to the agent.
[COMMAND_EXECUTION]: The skill executes local git commands to retrieve metadata about file modification times.
Evidence: scripts/scan_freshness.py and scripts/shared.py utilize subprocess.run() to execute git log and git rev-parse commands. These calls are implemented using argument lists rather than shell strings, which follows security best practices for command execution.

doc-freshness