extract-page-data
Fail
Audited by Gen Agent Trust Hub on Apr 14, 2026
Risk Level: HIGHCOMMAND_EXECUTIONEXTERNAL_DOWNLOADSDATA_EXFILTRATIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The script
scripts/lib/browser.mjsusesexecSyncto execute shell commands for installing theplaywrightpackage and the Chromium browser environment.\n- [EXTERNAL_DOWNLOADS]: The skill automatically downloads and installs the Playwright framework and Chromium browser from official registries and well-known mirrors during its initial execution.\n- [DATA_EXFILTRATION]: The Playwright browser is launched with the--disable-web-securityand--allow-file-access-from-filesflags. These settings disable the Same-Origin Policy (SOP) and permit the browser to access the local file system. If the agent is directed to a malicious website, that site could potentially read local files from the host machine or exfiltrate data from other domains the user is authenticated with.\n- [PROMPT_INJECTION]: The skill extracts content (Markdown, interactive elements, and design tokens) from external URLs, creating a surface for indirect prompt injection attacks.\n - Ingestion points: External webpage content is ingested via Playwright navigation in
scripts/extract.mjs.\n - Boundary markers: Absent. Extracted data is not wrapped in protective delimiters or accompanied by warnings to ignore embedded instructions.\n
- Capability inventory: The skill possesses capabilities for command execution, file system writes, and network navigation.\n
- Sanitization: There is no evidence of content sanitization or filtering of the extracted DOM content before it is processed by the agent.
Recommendations
- AI detected serious security threats
Audit Metadata