The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill instructs the agent to read and process external repository data that could contain malicious instructions.
Ingestion points: The agent reads repository instructions, git status, branch names, commit history, and merge request history (SKILL.md).
Boundary markers: The instructions include explicit guardrails requiring the agent to state its intended base/target branches before acting and to stay within previously granted authorization.
Capability inventory: The skill utilizes shell commands through git and the glab CLI tool.
Sanitization: No specific sanitization or filtering of commit messages or branch names is defined to prevent instruction processing from these sources.

git-flow