The Agent Skills Directory

[METADATA_POISONING]: The skill metadata (name, description, author) is consistent with the provided instructions and does not contain deceptive patterns or hidden commands.\n- [INDIRECT_PROMPT_INJECTION]: The skill's primary function is to process user-supplied prose, which represents a surface for indirect prompt injection. \n * Ingestion points: The agent is instructed to process and rewrite user-provided prose (SKILL.md).\n * Boundary markers: None specified in the instructions.\n * Capability inventory: The skill utilizes file system tools (Write, Edit) to modify text files.\n * Sanitization: No explicit sanitization or escaping is mentioned, though instructions mandate strict adherence to source facts to prevent hallucinations.\n- [UNVERIFIABLE_DEPENDENCIES_AND_REMOTE_CODE_EXECUTION]: No external packages, remote scripts, or unverified dependencies are requested or executed by this skill.

antislop-copywriting