The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is vulnerable to indirect prompt injection through the processing of git diffs.
Ingestion points: The agent reads the full <base>..HEAD diff to generate or update PR descriptions as described in SKILL.md. This diff content is untrusted external data.
Boundary markers: The instructions do not specify any delimiters or safety markers to differentiate between the agent's instructions and the content of the diff being analyzed.
Capability inventory: The skill utilizes the GitHub CLI (gh pr create and gh pr edit) to write and update pull request descriptions on remote repositories.
Sanitization: There are no instructions for the agent to sanitize, escape, or ignore potential instructions embedded within the code comments or commit messages in the diff.

write-pr