The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to process and follow instructions from external project files, creating a potential vector for malicious content.
Ingestion points: The agent is directed to read .github/pull_request_template.md and GLOSSARY.md from the target repository to structure its output.
Boundary markers: There are no explicit delimiters or safety instructions provided to the agent to treat the contents of the pull request template as untrusted data.
Capability inventory: The skill is restricted to generating text within the user's conversation; it does not include instructions for shell command execution, network access, or unauthorized file system access.
Sanitization: The skill does not validate or sanitize the content of the template or glossary files before processing them.
[NO_CODE]: The skill consists entirely of markdown instructions and YAML metadata. It includes no scripts, binaries, or other executable components, which significantly limits the potential for direct security exploits.

pr