inboundfy-extrair-faq
Pass
Audited by Gen Agent Trust Hub on Oct 12, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes external data from "bruto.md" and "processado.md" to extract FAQs, which creates an ingestion surface for indirect prompt injection if those source files contain malicious instructions.
- Ingestion points: Processes content from "bruto.md" and "processado.md" as defined in the 'Entrada esperada' section of SKILL.md.
- Boundary markers: The skill does not define explicit delimiters to isolate external content, although it includes instructional constraints to use only provided sources.
- Capability inventory: The skill performs local file read and write operations (creating "faq.md"). It does not define tools with network access, shell execution, or high-privilege capabilities in its flow.
- Sanitization: The instructions explicitly forbid the creation of facts not found in the source material and require specific citations (e.g., "Origem: bruto.md#trecho"), which provides a logical check against data manipulation.
Audit Metadata