The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted writing samples and profile documents to extract stylistic patterns.
Ingestion points: The skill reads user-provided writing samples (such as social media posts, transcripts, and DMs) and a brand-profile.md file to perform its core analysis.
Boundary markers: The instructions do not specify the use of delimiters or specific tags to separate untrusted sample content from the agent's internal reasoning or instructions.
Capability inventory: The skill is restricted to reading and writing local text files (voice.md, brand-profile.md) and generating text based on prompts. It lacks network access or the ability to execute shell commands.
Sanitization: The skill lacks explicit instructions to sanitize or escape the content of the writing samples, which could allow embedded instructions in the samples to influence the agent's behavior during the analysis or validation phases.

voice-builder