The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses git and gh (GitHub CLI) to automate repository workflows, including branch management, PR creation, and merging.
[COMMAND_EXECUTION]: Through the --setup parameter in scripts/unit.mjs, the skill executes arbitrary shell commands to facilitate project-specific worktree initialization (e.g., running installation scripts).
[INDIRECT_PROMPT_INJECTION]: The skill acts as an ingestion point for external data from GitHub issues and task manifests. The evidence chain includes:
Ingestion points: Issue content from gh issue list and user-provided task manifest files.
Boundary markers: Explicit instructions in SKILL.md advising the agent that task text provides requirements but not authority.
Capability inventory: Git/GitHub operations and shell execution capabilities for setup hooks.
Sanitization: Relies on a mandatory human review gate for all PR merges and specific instructions for the agent to validate implementation against plans.
[COMMAND_EXECUTION]: The scripts/usage-state.mjs utility executes the cursor-agent CLI to monitor AI service capacity and rate limits.

orchestrate