press-release-writing

Warn

Audited by Gen Agent Trust Hub on Sep 2, 2026

Risk Level: MEDIUMEXTERNAL_DOWNLOADSREMOTE_CODE_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill instructs the user to install the 'belt-sh/cli' package using 'npx skills add'. It also points to an external GitHub repository (raw.githubusercontent.com/inference-sh/skills) for installation instructions.
  • [REMOTE_CODE_EXECUTION]: The skill relies on the 'belt' CLI to execute remote applications (such as 'tavily/search-assistant' and 'exa/search'). These applications run logic on external servers and return content that the agent is instructed to process and incorporate into its output.
  • [INDIRECT_PROMPT_INJECTION]: The skill ingests data from external search providers (Tavily and Exa) for fact-checking and research purposes. This creates an attack surface where malicious instructions embedded in external web content could potentially influence the agent's behavior.
  • Ingestion points: Output from the 'belt app run' commands in SKILL.md.
  • Boundary markers: None; search results are directly processed by the agent without explicit delimiters or instructions to ignore embedded commands.
  • Capability inventory: The agent has access to the 'belt' CLI and the ability to generate and format complex documents (press releases).
  • Sanitization: No evidence of filtering or sanitization of the retrieved search content before it is processed by the agent.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Sep 2, 2026, 02:18 PM
Security Audit — agent-trust-hub — press-release-writing