The Agent Skills Directory

[PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface by interpolating untrusted repository data into subagent contexts.
Ingestion points: The agent reads project-specific data including milestone acceptance criteria, scope items, release checklists, and bug reports using the Read, Glob, and Grep tools.
Boundary markers: The instructions do not define clear delimiters or instruct subagents to ignore potential instructions embedded within the processed documentation or data files.
Capability inventory: The skill possesses high-privilege capabilities including the Bash tool for shell execution, Write/Edit for file modification, and the Task tool for spawning further subagents, creating a significant impact if the agent is manipulated.
Sanitization: There is no evidence of sanitization or validation for the data ingested from the repository before it is used to influence the release pipeline or command execution.

team-release