claude-video-screenshot
Pass
Audited by Gen Agent Trust Hub on Sep 20, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADSCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to visit and capture content from arbitrary external URLs provided by the user or as part of a batch process. This content is untrusted and could contain malicious instructions meant to influence the agent's context or behavior during the session.
- Ingestion points: Playwright browser engine loading external URLs via the
--urland--urlsparameters inSKILL.md. - Boundary markers: No specific delimitation or instruction to ignore embedded content is mentioned for the ingested web data.
- Capability inventory: The skill utilizes
Bash,Read, andWritetools to execute scripts and manage output files. - Sanitization: No sanitization of the external web content is described prior to processing.
- [EXTERNAL_DOWNLOADS]: The skill downloads the Playwright library and necessary Chromium browser binaries from official package registries and vendor distribution points.
- Evidence:
pip install playwright && playwright install --with-deps chromiuminSKILL.md. - [COMMAND_EXECUTION]: The skill executes local Python scripts and system utilities like FFmpeg via the Bash tool to manage web interactions and generate video animations.
- Evidence: Usage of
python3 scripts/web_capture.pyand documentation offfmpegcommand strings inSKILL.md.
Audit Metadata