The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from the repository, such as file contents, diffs, and issue descriptions, which could contain malicious instructions designed to influence the agent's behavior during the shipping process.
Ingestion points: SKILL.md instructs the agent to run git status and read every changed file, scan diffs for secrets, and link/close issues based on their content.
Boundary markers: The instructions do not define specific delimiters or "ignore instructions" warnings for the content read from files or issues.
Capability inventory: The agent can execute git commands (commit, push, merge, branch deletion), open URLs to verify deployment, and interact with issue trackers.
Sanitization: There are no specific sanitization or filtering steps mentioned for the external content before it is processed by the agent.

ship