skills/akillness/jeo-skills/compresso/Gen Agent Trust Hub

compresso

Pass

Audited by Gen Agent Trust Hub on Nov 3, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSREMOTE_CODE_EXECUTIONCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill instructs users to download binaries (AppImage, DMG, MSI) and source code from the "codeforreal1/compressO" repository on GitHub.
  • [REMOTE_CODE_EXECUTION]: The installation instructions include fetching the skill via "npx" from a third-party GitHub repository ("akillness/jeo-skills") and installing software through a custom Homebrew tap ("codeforreal1/tap/compresso").
  • [COMMAND_EXECUTION]: The documentation encourages the execution of external binaries after modifying permissions ("chmod +x") and building the application from source using "pnpm install", which downloads and executes third-party Node.js dependencies.
  • [INDIRECT_PROMPT_INJECTION]:
  • Ingestion points: The skill is designed to process external video and image files and their associated metadata (SKILL.md).
  • Boundary markers: No explicit instructions are present to ensure that the agent ignores instructions that could be embedded in the processed files' metadata or subtitles.
  • Capability inventory: The skill leverages powerful system tools such as FFmpeg and Tauri (SKILL.md).
  • Sanitization: There is no mention of sanitizing or validating the content or metadata of the files before they are processed by the underlying tools.
Audit Metadata
Risk Level
SAFE
Analyzed
Nov 3, 2026, 05:32 AM