The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted external data (X posts, emails, DMs, site copy) to extract stylistic patterns, which could contain hidden instructions.
Ingestion points: Defined in SKILL.md under the 'Collection Workflow' section, which instructs the agent to gather 5 to 20 representative samples, pull recent posts via x-api, and include landing page copy.
Boundary markers: The instructions lack explicit boundary markers or delimiters to separate the source material from the agent's instructions, and there are no 'ignore embedded instructions' warnings.
Capability inventory: While this skill primarily generates a profile, it is designed to be used downstream by skills like content-engine, crosspost, and lead-intelligence, which typically possess network access and file-writing capabilities.
Sanitization: No sanitization, filtering, or validation logic is specified for the ingested content before it is processed by the agent.

brand-voice