cawcut-workflow
Pass
Audited by Gen Agent Trust Hub on Oct 29, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill manages its execution environment by checking for updates using the
cawcut upgradecommand and providing instructions to install the@ubnt/cawcutpackage from the NPM registry if the CLI tool is not found. - [COMMAND_EXECUTION]: The skill extensively uses the
Bashtool to execute thecawcutCLI for operations including authentication, media uploads, workflow orchestration, and task monitoring. - [INDIRECT_PROMPT_INJECTION]: The skill possesses an indirect injection surface by scraping content from product URLs to generate media prompts and scripts.
- Ingestion points: In
references/product.md, the skill scrapes product pages to extract 'brief' text and images. - Boundary markers: No specific delimiters or safety instructions are mentioned for handling the extracted text during prompt interpolation.
- Capability inventory: The agent has access to the
Bashtool and thecawcutCLI to perform network and file operations. - Sanitization: There is no evidence of sanitization or validation performed on the scraped content before it is passed to the LLM for script writing or prompt generation.
Audit Metadata