The Agent Skills Directory

[SAFE]: No security issues were detected. The skill consists entirely of natural language instructions for the agent.
[SAFE]: The skill lacks executable scripts, remote dependencies, or command execution, minimizing the attack surface.
[SAFE]: Explicit guardrails are provided in the 'Pitfalls / Guardrails' section to prevent the agent from performing unauthorized tasks like scanning, translation, or modification.
[SAFE]: No evidence of prompt injection, data exfiltration, or obfuscation was found in the content or metadata.
[SAFE]: Although the skill processes external data (code, documents, datasets), it lacks high-risk capabilities such as network access or shell execution, mitigating the risk of indirect prompt injection.
Ingestion points: User-provided documents and datasets (SKILL.md).
Boundary markers: The procedure includes a step to restate targets and boundaries, but no explicit data delimiters are defined.
Capability inventory: No subprocess calls, network operations, or file-system writing capabilities are requested or used.
Sanitization: No explicit sanitization or filtering of external content is mentioned.

faerie-fire