The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses git and gh (GitHub CLI) commands to manage the development cycle, including creating git worktrees, checking out branches, merging pull requests, and managing labels.
[EXTERNAL_DOWNLOADS]: The skill automatically performs package installation using bun install or npm install when setting up its worktree. This is a standard development workflow that downloads dependencies from external registries.
[PROMPT_INJECTION]: The skill is vulnerable to indirect prompt injection (Category 8) because it reads and acts upon content provided by external users in GitHub issues and PR comments.
Ingestion points: Untrusted data enters the agent context through pull request descriptions (gh pr view), issue bodies (gh issue view), and API responses for PR comments.
Boundary markers: The skill does not implement boundary markers or instructions to the agent to ignore potentially malicious directions embedded in the fetched issue or PR text.
Capability inventory: The skill possesses significant capabilities, including the ability to execute shell commands, write to the local filesystem (via implementation tools), and modify remote repository states (merging/closing PRs).
Sanitization: There is no evidence of sanitization, validation, or filtering of the content retrieved from GitHub before it is processed by the agent.

supervisor