The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill ingests 'gate name' and 'pass criteria' to define its verification loop. While this is the intended functionality, it creates a surface for indirect instructions if the criteria are sourced from untrusted data.
Ingestion points: SKILL.md Setup section (Gate name, Pass criteria fields).
Boundary markers: Absent; the skill does not define specific delimiters for these inputs.
Capability inventory: Uses Read, Write, Bash, and Agent tools to perform verification and repairs.
Sanitization: Not present; the skill assumes the agent will interpret the criteria directly.

convergence-loop