cueframe-brand-demo
Pass
Audited by Gen Agent Trust Hub on Nov 11, 2026
Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSDATA_EXFILTRATIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill instructs the agent to run local automation scripts, specifically Playwright walkthroughs and a native screen recorder, to capture the application's UI.
- [EXTERNAL_DOWNLOADS]: The documentation mentions installing the vendor's tools via 'npx -y cueframe@0.12 install', which downloads and executes code from the npm registry.
- [DATA_EXFILTRATION]: To enable cloud rendering, the skill requires local assets (such as logos found in 'public/' or 'assets/') to be uploaded or hosted at public URLs for the CueFrame engine to fetch.
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from the user's local repository, creating an attack surface where malicious instructions in project files could influence agent behavior.
- Ingestion points: Reads project files including 'tailwind.config.*', 'design.md', 'README', and landing page copy in 'SKILL.md'.
- Boundary markers: No specific delimiters or 'ignore embedded instructions' warnings are provided for the content read from the codebase.
- Capability inventory: The skill can call 'create_brand_kit', 'import_media', 'apply_composition', 'compose', and 'create_render' APIs, as well as execute local CLI tools.
- Sanitization: There is no mention of sanitizing or validating the text extracted from READMEs or code before it is used to synthesize brand parameters or voice guidelines.
Audit Metadata