frevana-space-cms
Pass
Audited by Gen Agent Trust Hub on Oct 9, 2026
Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill invokes a bundled bash script (
publish_file.sh) and embedded Python code blocks to coordinate API calls, handle responses, and parse HTML metadata. - [EXTERNAL_DOWNLOADS]: The skill performs network operations to
api.frevana.comand communicates with object storage endpoints to upload and publish user content. These requests are directed to the vendor's official infrastructure. - [INDIRECT_PROMPT_INJECTION]: The skill has an ingestion surface for untrusted data as it processes local HTML files provided by the user or agent to extract titles and perform uploads.
- Ingestion points: The script
scripts/publish_file.shreads the contents of local.htmlfiles provided as input. - Boundary markers: No specific boundary markers or instructions are utilized to isolate the parsed file content from the agent's instructional context.
- Capability inventory: The skill combines file-reading capabilities with network access (
curl) for data exfiltration (uploads). - Sanitization: Metadata extraction logic uses a structured
HTMLParserand regular expressions to clean and normalize titles before inclusion in API requests.
Audit Metadata