browser-pair
Pass
Audited by Gen Agent Trust Hub on Apr 12, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [PROMPT_INJECTION]: The skill facilitates an indirect prompt injection surface by reading and processing content from external web pages during headed browser sessions. \n- Ingestion points: The skill uses
mcp__playwright-headed__browser_snapshotto read the state of the current web page into the agent's context. \n- Boundary markers: There are no explicit instructions to use delimiters or ignore embedded instructions within the page data, although the workflow mandates a 'Wait for user direction' step which acts as a manual boundary. \n- Capability inventory: The agent has access tomcp__playwright-headed__browser_click,mcp__playwright-headed__browser_type, andmcp__playwright-headed__browser_navigate, as well as the ability to read and edit project files. \n- Sanitization: No sanitization or validation of the ingested page content is specified before the agent processes it.
Audit Metadata