The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process untrusted external data, specifically transcripts, notes, and audio from user validation sessions.
Ingestion points: In Fase 1 and Fase 2, the skill is instructed to extract findings and decisions from session recordings, transcriptions, and notes.
Boundary markers: Absent. There are no explicit instructions or delimiters provided to ensure the agent ignores potential commands or instructional text embedded within the validation data.
Capability inventory: The skill utilizes tools with significant impact on the project environment, including transition_ficha (status changes), eliminar_ficha (record deletion), and crear_ticket (task creation).
Sanitization: Absent. The instructions do not specify any validation or sanitization steps to filter out malicious instructions hidden within the user-provided content.

wachi-validacion-usuario