The Agent Skills Directory

[COMMAND_EXECUTION]: The skill includes several Python snippets designed for verification and instructs the agent to execute them using an evaluation tool. Specifically, references/source-map.md contains code to validate documentation frontmatter using pathlib and PyYAML, while references/verify.md provides a script for checking internal link consistency across the docs/ directory. The skill also executes a series of shell commands, including git rev-parse, git add, and git commit, as part of its core documentation workflow. Furthermore, it suggests using shell commands like find to check for CLI subcommand parity within the source tree.
[PROMPT_INJECTION]: The skill exhibits a significant attack surface for indirect prompt injection because it parses and documents the entire codebase of the repository it is running in.
Ingestion points: The skill reads from various directories including cmd/, internal/, pkg/, and src/, as well as configuration files like go.mod, package.json, or Cargo.toml and existing project documentation.
Boundary markers: No specific delimiters or safety instructions are provided to the agent to distinguish between its operational instructions and the potentially malicious data contained within the source code files being read.
Capability inventory: The agent possesses high-impact capabilities, including shell access for Git operations and the ability to execute arbitrary Python code via the provided scripts.
Sanitization: The skill lacks mechanisms to sanitize or validate the content extracted from the source code, allowing data from the repository to directly influence the agent's logic and outputs.

write-docs