The Agent Skills Directory

Indirect Prompt Injection Surface: The skill defines procedures for ingesting external YAML payloads. While ingesting external data is a potential risk, the skill manages this via rigorous sanitization and validation. 1. Ingestion points: Data enters through user-supplied YAML artifacts like outcome hypotheses and feasibility studies as described in the reference documentation. 2. Boundary markers: The agent is instructed to validate payloads before ingestion and reject unstructured interpretations of failed payloads. 3. Capability inventory: The skill uses file-reading and template-generation capabilities to manage requirements documents. 4. Sanitization: The skill enforces strict schema validation, version checks, and field-level verification. These measures effectively constrain untrusted data within its defined scope.
Data Integrity Practices: The lifecycle incorporates integrity checks, such as SHA-256 hashing of finalized artifacts and tracking document lineage. These features ensure the authenticity and consistency of the requirements documentation across its phases.
Capabilities Scope: The skill functions by processing markdown templates and structured metadata. It does not require network access, handle sensitive credentials, or execute dynamic code, maintaining a restricted execution environment.

requirements-author