The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to process untrusted external inputs like architectural designs and changesets to generate a task manifest that includes executable verification commands.
Ingestion points: Ingests architectural directions and contracts in SKILL.md Execution Protocol.
Boundary markers: No explicit delimiters are provided to distinguish between ingested data and task-generation instructions.
Capability inventory: The skill directs output to tools such as /forge and /changeset for execution and file manipulation.
Sanitization: No sanitization logic is present for the verification criteria field.
[NO_CODE]: The skill contains only natural language instructions and structural templates, with no executable code or external library dependencies.

split-tasks