The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process external data in the form of specifications and issue tracker tickets. This creates a surface for indirect prompt injection if these external sources contain malicious instructions.
Ingestion points: Instructions in SKILL.md specify reading the "spec and tickets" to understand the task graph and drive the implementation workflow.
Boundary markers: The instructions do not define clear delimiters or "ignore embedded instructions" warnings when processing these external inputs.
Capability inventory: The skill coordinates complex git workflows including branch management, worktree creation, merging, and PR management. It also triggers further "implementer subagents" and invokes testing tools like tdd.
Sanitization: There is no explicit sanitization, validation, or filtering logic described for the content of the specs or tickets before they are used to direct the agent's actions.

implement-spec