The Agent Skills Directory

[PROMPT_INJECTION]: The skill processes potentially untrusted external content, such as bio drafts and spokesperson answers, which serves as an ingestion point for indirect prompt injection.
Ingestion points: User-provided bio drafts, spokesperson answers, and pasted canon excerpts.
Boundary markers: The instructions explicitly state to treat these inputs as untrusted and to ignore any instructions embedded within them.
Capability inventory: The skill has the ability to write to the memory/ directory and execute the local script registry-events.py to propose new data entries.
Sanitization: The skill includes a step to sweep output against a banned-vocabulary list defined in the skill contract.
[COMMAND_EXECUTION]: The skill is configured to execute a local script named registry-events.py to handle authorized data proposals to the narrative registry. This is a functional requirement for maintaining the agent's memory systems and is limited to internal operations.

narrative-enablement-kit