The Agent Skills Directory

[PROMPT_INJECTION]: The skill is vulnerable to indirect prompt injection because it ingests and processes content from external codebase files which could contain hidden malicious instructions.
Ingestion points: Reads local project configuration files (e.g., package.json, Cargo.toml, requirements.txt), existing README.md files, and source code files.
Boundary markers: None identified; the skill does not explicitly use delimiters or instructions to ignore embedded commands within the files it analyzes.
Capability inventory: Includes file system access (listing and reading files), execution of git commands, and the ability to write/modify the README.md file.
Sanitization: No evidence of sanitization or validation of the content extracted from the codebase before it is used in the generation prompt.
[COMMAND_EXECUTION]: The skill utilizes system-level commands to perform its analysis.
Evidence: Specifically references running git diff and git log to detect codebase changes and understand project history.

readme-generator