inboundfy-estrategia
Pass
Audited by Gen Agent Trust Hub on Oct 12, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and analyze potentially untrusted external data during its operation.
- Ingestion points: In 'references/etapas/00-briefing-cliente.md', it receives kickoff conversations and informal briefs. In 'references/etapas/01-pesquisa-mercado.md', it researches and analyzes content published by competitors.
- Boundary markers: The skill implements a structured execution architecture with 'Preflight' and 'Validation' stages, and it explicitly requires separating facts from interpretations in its output.
- Capability inventory: Across its scripts, the skill is authorized to write Markdown files to specific directories (e.g., '/01-plano/') and hand off tasks to other internal skills within the system.
- Sanitization: No explicit sanitization or filtering of the ingested natural language content is described, which is typical for instruction-based skills.
Audit Metadata