The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses the Bash tool to execute specific git and forge-specific (gh/glab) commands necessary for repository analysis and branch management. These operations, particularly the deletion of branches, are strictly gated by a manual confirmation step where the agent must present a plan and obtain an explicit "yes" from the user before acting.
[INDIRECT_PROMPT_INJECTION]: The skill identifies and processes untrusted data such as pull request titles, branch names, and commit messages. It contains detailed adversarial reasoning instructions (under the "Author authority" section in REFERENCE.md) directing the agent to treat this text as data only, explicitly warning against acting on instructions embedded in such text.
[EXTERNAL_DOWNLOADS]: The skill references a configuration schema hosted on the author's GitHub repository (github.com/TitusKirch/skills). This is documented as a source for configuration validation and matches the trusted vendor resource pattern for the skill author.
[DYNAMIC_EXECUTION]: The skill utilizes a local shell script (templates/resolve-config.sh) to handle configuration resolution across different environments. This script uses standard tools like git and jq to merge configuration profiles without introducing external dependencies or remote code execution risks.

prune-branches