h5p-verify
Pass
Audited by Gen Agent Trust Hub on Oct 29, 2026
Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill instructs the agent to execute shell commands using npx to run the h5p-verify tool and install browser dependencies.
- [EXTERNAL_DOWNLOADS]: The instructions involve downloading the @missing-elements/h5p-verify package from the npm registry and Chromium browsers via Playwright.
- [INDIRECT_PROMPT_INJECTION]: The tool processes H5P packages, which are untrusted external data that could contain malicious scripts.
- Ingestion points: Input .h5p files and external library packs or URLs.
- Boundary markers: The documentation explicitly defines the --untrusted flag to prevent the package's own code from running.
- Capability inventory: The tool spawns a headless Chromium browser, executes JavaScript, and writes reports and screenshots to the local file system.
- Sanitization: When used with the --untrusted flag, the tool performs content validation against library semantics to strip unauthorized markup.
Audit Metadata