The Agent Skills Directory

[COMMAND_EXECUTION]: The skill instructs the agent to perform repository management and CI/CD operations using standard CLI tools including git, gh (GitHub CLI), and glab (GitLab CLI).
[COMMAND_EXECUTION]: The skill defines a structured routing mechanism that generates and executes commands for other agent skills, such as ao spawn and clickup-api, based on task descriptions.
[DATA_EXFILTRATION]: The instructions involve reading from and writing to configuration and state directories like ~/.claude/ and managing repository paths within ~/Sync/ to verify repo identity and avoid 'supergit' traps.
[PROMPT_INJECTION]: The skill acts as an intermediary for arbitrary engineering task descriptions (ingestion point: 'Clarify' stage), creating an indirect prompt injection surface. The skill mitigates this through structured 'Decision Card' outputs and mandatory pre-flight validation gates. Evidence chain:
Ingestion points: Task descriptions provided during Stage 1 (Clarify).
Boundary markers: None explicitly defined for the input task description.
Capability inventory: ao spawn, git, gh, glab, clickup-api (documented in SKILL.md).
Sanitization: Relies on mandatory pre-flight gates and 'codex-pre-review' for destructive operations.

agent-runbook