The Agent Skills Directory

[PROMPT_INJECTION]: The skill is vulnerable to indirect prompt injection through the processing of untrusted repository data.
Ingestion points: Commands such as gk ai explain, gk ai changelog, and gk ai commit in SKILL.md ingest git diffs, commit history, and branch content.
Boundary markers: The skill does not define boundary markers or provide instructions to the agent to disregard instructions embedded within the processed git data.
Capability inventory: The skill possesses capabilities to write to repositories, push code, and create pull requests using commands like gk work push and gk ai pr create.
Sanitization: No sanitization or validation of repository content is performed before it is passed to the AI-driven commands.
[COMMAND_EXECUTION]: The skill facilitates the execution of the gk (GitKraken) command-line interface. It enables the agent to perform repository management tasks, including pushing code and managing work items, which are triggered based on the agent's interpretation of user goals.

gitkraken-cli