press-release-writing
Pass
Audited by Gen Agent Trust Hub on Nov 14, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADSCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes user-supplied information to draft press releases and execute research queries via external search tools.
- Ingestion points: User-provided data for press release topics and specific queries passed to
belt app runinSKILL.md. - Boundary markers: Absent; inputs are directly used in tool parameters.
- Capability inventory: Network access for research via
tavily/search-assistantandexa/searchtools. - Sanitization: No input validation or sanitization is demonstrated in the provided instructions.
- [EXTERNAL_DOWNLOADS]: Instructions reference downloading and installing the
belt-sh/cliand additional skills from theinference-shGitHub organization. These represent well-known services and trusted tooling within the agent ecosystem. - [COMMAND_EXECUTION]: The skill utilizes the
beltCLI tool to perform search and fact-checking operations as part of its core functionality.
Audit Metadata