The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes multiple shell commands to perform its tasks.
Uses git diff and git status to analyze code changes.
Invokes local tools and aliases like ti (issue tracker) and tpr (PR review) using zsh -ic to load the user's interactive shell environment.
Runs mise run blc and mise run pw-login for repository-specific loops and authentication.
[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data which could contain malicious instructions.
Ingestion points: Reads content from branch issue descriptions (ti <n>), PR review comments (tpr <pr>), and a repository-level AGENTS.md file (SKILL.md, Step 1 and Step 4).
Boundary markers: No specific delimiters or safety instructions are defined to separate the ingested data from the agent's internal logic.
Capability inventory: The skill has the ability to execute shell commands, stage files via git, and trigger other agent skills (code-review, caveman-commit).
Sanitization: The instructions do not specify any sanitization, validation, or escaping for the external data fetched from the issue tracker or PR comments.

work-wrap-up