The Agent Skills Directory

[DATA_EXFILTRATION]: The skill transmits integration summaries and user feedback to the vendor's domain (api.liveavatar.com). This process is transparent, requires explicit user consent, and allows the user to review and edit all data before transmission.- [INDIRECT_PROMPT_INJECTION]: The skill ingests raw user feedback for submission to a remote API, which represents a potential injection surface.
Ingestion points: User input in the additional_feedback field defined in SKILL.md.
Boundary markers: None present in the curl payload construction.
Capability inventory: Network POST request capabilities using curl as defined in SKILL.md.
Sanitization: Instructions explicitly direct the agent to use the user's exact words verbatim without rephrasing or summarization.

liveavatar-feedback