The Agent Skills Directory

[Indirect Prompt Injection] (LOW): The skill defines a protocol for delegating work to sub-agents using TASK and CONTEXT placeholders, which are typically populated with untrusted data. 1. Ingestion points: TASK and CONTEXT fields defined in the Spawn Protocol section of SKILL.md. 2. Boundary markers: Employs simple text delimiters such as === WORKER AGENT ===. 3. Capability inventory: Instructions describe spawning and monitoring agents for complex tasks, architectural decisions, and performance-critical code. 4. Sanitization: No sanitization or validation logic is defined for the interpolated content.
[No Code] (INFO): The provided skill contains only markdown documentation and protocol definitions; it does not include executable code, environment configurations, or network requests.

team-coordination