paw-mkt-cro
Pass
Audited by Gen Agent Trust Hub on Sep 15, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill downloads and installs the
agent-browserextension from Vercel Labs' official GitHub repository usingnpxto facilitate live website auditing. - [COMMAND_EXECUTION]: The skill executes local shell scripts for environment setup and tool discovery, such as
./skills/paw-mkt-setup/assets/scripts/tool-discovery.sh, to verify the presence of required runtimes and CLI tools. - [INDIRECT_PROMPT_INJECTION]: The skill has an attack surface for indirect prompt injection as it ingests and processes untrusted content from external websites during live audits.
- Ingestion points: The
agent-browser get text bodycommand inreferences/shared-patterns.mdis used to retrieve the visible text of any URL provided by the user. - Boundary markers: The instructions do not specify the use of clear delimiters or guardrail instructions to ensure the agent ignores potential commands embedded in the fetched web content.
- Capability inventory: The skill includes capabilities to write audit reports to the filesystem, execute browser interactions, and run local setup scripts.
- Sanitization: There is no evidence of filtering, escaping, or sanitizing the retrieved website content before it is processed by the agent.
Audit Metadata