press-release-writing
Pass
Audited by Gen Agent Trust Hub on Aug 28, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill instructs the agent to use external search applications (
tavily/search-assistant,exa/search) to gather facts and context for writing. - Ingestion points: Results from external API queries are brought into the agent's active context via the
beltCLI (SKILL.md). - Boundary markers: The instructions do not define boundary markers or explicit safety prompts to ensure the agent ignores any malicious instructions potentially embedded in search results.
- Capability inventory: The skill is configured with
Bash(belt *)permissions, allowing it to execute CLI commands (SKILL.md). - Sanitization: No sanitization or data validation steps are specified for the search output before it is processed by the agent.
Audit Metadata