skills/pixazo-ai/skills/cosmos/Gen Agent Trust Hub

cosmos

Pass

Audited by Gen Agent Trust Hub on Sep 22, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTIONREMOTE_CODE_EXECUTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill retrieves model metadata and polling instructions from the official vendor domain at https://www.pixazo.ai/models/cosmos.md.
  • [INDIRECT_PROMPT_INJECTION]: The skill is instructed to dynamically load parameters from a remote markdown file. While this represents a surface for indirect instructions, the source is limited to the official vendor documentation.
  • Ingestion points: documentation fetch referenced in SKILL.md
  • Boundary markers: Absent
  • Capability inventory: network requests (curl, requests, fetch), local file writing, chmod, and package execution (npx)
  • Sanitization: None.
  • [COMMAND_EXECUTION]: The instructions include saving the user's API key to a local configuration file and setting appropriate file system permissions using chmod 600.
  • [REMOTE_CODE_EXECUTION]: The skill documentation suggests using npx to install and run additional skills from the vendor's official repository.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 22, 2026, 09:01 PM