The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted external data (such as Instagram captions, carousels, and scripts) to perform validation, which introduces a surface for indirect prompt injection attacks.
Ingestion points: The skill ingests an 'Asset candidato', 'brief', and 'fontes' as defined in the 'Entrada esperada' section of SKILL.md.
Boundary markers: While the skill utilizes a structured output template in REFERENCIA.md, it lacks explicit delimiting or safety instructions to the agent to ignore instructions that might be embedded within the asset content.
Capability inventory: The skill reads various project context files (e.g., .inboundfy/fontes-projeto.md, context/marca-voz.md) and generates reports, potentially invoking the inboundfy-base-validador tool.
Sanitization: There are no mentioned processes for sanitizing or escaping the content of the external assets before the agent processes them.
[COMMAND_EXECUTION]: The instructions in SKILL.md direct the agent to suggest the execution of the inboundfy-setup command to the user if specific configuration files are not found during the preflight check.

inboundfy-validador-instagram