css-animation-skill
Fail
Audited by Gen Agent Trust Hub on Jun 25, 2026
Risk Level: HIGHEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONREMOTE_CODE_EXECUTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill instructions direct the agent to fetch and install external code from 'https://github.com/neonwatty/css-animation-skill'. This facilitates the introduction of third-party content into the execution environment.\n- [REMOTE_CODE_EXECUTION]: The use of 'npx skills add' triggers the downloading and execution of scripts from an unverified source. This 'stub' pattern allows a seemingly simple skill to transition into a complex, remotely-hosted skill at runtime.\n- [COMMAND_EXECUTION]: The workflow relies on shell commands like 'grep' and directory existence checks ('[ -e ... ]') to determine its internal state and installation status.\n- [INDIRECT_PROMPT_INJECTION]: The skill is designed to scrape 'live apps' via MCP tools to extract design languages. This ingestion of untrusted external data could allow a malicious website to influence the agent's behavior.\n
- Ingestion points: Claude-in-Chrome MCP scraping of live web application content.\n
- Boundary markers: None detected in the instructions to prevent the agent from obeying instructions found within the scraped data.\n
- Capability inventory: Execution of shell commands ('npx', 'grep') and file system interaction.\n
- Sanitization: No evidence of sanitization or validation of the scraped visual/textual data.
Recommendations
- AI detected serious security threats
Audit Metadata