The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted user input, such as drafts for tweets, threads, and past post samples, to perform rewrites and build voice profiles. This creates a surface for indirect prompt injection where instructions embedded in the input could theoretically influence the agent's behavior.
Ingestion points: Input text processed in SKILL.md and user-provided samples used in sub-skills/voice-profile.md.
Boundary markers: The instructions do not define specific delimiters or warnings to ignore instructions embedded within the processed content.
Capability inventory: The skill performs text generation and invokes image generation tools via lib.illustrate.
Sanitization: No explicit sanitization or input filtering mechanisms are described for the processed text.
[EXTERNAL_DOWNLOADS]: The skill integrates with the Pixfaro image generation service to create and host visual content for social media posts.
Evidence: Found in sub-skills/illustration.md, which details the use of PIXFARO_TOKEN for automated image generation and attachment. This is a functional feature for creating quote-cards and feed images.

x-humanizer