The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes various git commands (git log, git diff, git push) and GitHub CLI commands (gh pr create, gh pr checks, gh run view) to manage repository workflows. These are standard operations for the stated purpose of the skill.
[DATA_EXFILTRATION]: The skill transmits local repository metadata, commit history, and diffs to GitHub's official services. This behavior is expected for a pull request automation tool and targets a well-known, trusted service.
[INDIRECT_PROMPT_INJECTION]: The skill processes external data that could potentially be controlled by an attacker, specifically CI logs and git diffs.
Ingestion points: Reads output from git diff, git log, and gh run view --log-failed (SKILL.md, Step 1 and Step 8).
Boundary markers: None present for the logs being processed by subagents.
Capability inventory: The agent can perform git push and gh pr create, allowing it to modify the repository (SKILL.md, Step 6 and Step 9).
Sanitization: No specific sanitization or filtering of the log content is performed before it is passed to subagents for analysis.

create-pr