press-release-writing
Warn
Audited by Gen Agent Trust Hub on Sep 2, 2026
Risk Level: MEDIUMEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill instructs users to install a CLI tool via
npx skills add belt-sh/cliand provides links to external installation scripts athttps://raw.githubusercontent.com/inference-sh/skills/refs/heads/main/cli-install.md. These resources originate from an external source not recognized as a well-known service or trusted organization. - [COMMAND_EXECUTION]: The skill utilizes the
Bashtool to runbeltcommands, includingbelt loginandbelt app run. These commands execute logic and remote applications from theinference.shplatform. - [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data retrieved from external tools (search results) and user-supplied content without explicit boundary markers or sanitization.
- Ingestion points: Research data entering via
belt app runinSKILL.md. - Boundary markers: Absent; external content is interpolated directly into the writing workflow.
- Capability inventory:
Bash(belt *)capability used for research and tool execution. - Sanitization: Absent; no filtering or escaping of external search results or user input is performed before generation.
Audit Metadata