The Agent Skills Directory

[COMMAND_EXECUTION]: The skill performs local command execution through standard Git operations (git diff, git show) and a repository-specific tool (./.planr/tooling/planr). These commands are used to verify code changes and are constrained to the local repository environment.
[PROMPT_INJECTION]: The skill processes untrusted repository data, including plan files and code diffs, which creates a surface for indirect prompt injection. The instructions mitigate this risk by mandating strict scope isolation and verification gates to ensure the agent maintains control over the review workflow.
[DATA_EXFILTRATION]: The skill accesses project-related files and Git metadata to perform its functions. No network communication patterns or external transmission methods were identified.

planr-review