chrome-devtools
Pass
Audited by Gen Agent Trust Hub on Nov 12, 2026
Risk Level: SAFECOMMAND_EXECUTIONDYNAMIC_EXECUTIONPRIVILEGE_ESCALATIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [PRIVILEGE_ESCALATION]: The
scripts/install-deps.shscript executes package manager commands withsudoto install system libraries necessary for running Chrome/Chromium on Linux environments. - [COMMAND_EXECUTION]: The installation process involves executing shell scripts (
install.sh,install-deps.sh) that interact with the host operating system's package management tools. - [DYNAMIC_EXECUTION]: The
scripts/evaluate.jstool utilizeseval()to execute user-provided JavaScript strings within the browser's page context. While intended for debugging, this constitutes a dynamic execution sink. - [INDIRECT_PROMPT_INJECTION]: The skill ingests data from external websites, which could potentially contain malicious instructions aimed at the agent.
- Ingestion points: Untrusted data enters the agent context through
snapshot.js(extracting DOM elements),console.js(monitoring logs),network.js(tracking HTTP traffic), andevaluate.js(executing scripts on pages). - Boundary markers: The scripts do not implement specific delimiters or warnings to prevent the agent from interpreting ingested web content as instructions.
- Capability inventory: The skill possesses powerful capabilities including browser interaction (clicking, filling forms), file system access (writing screenshots), and network monitoring.
- Sanitization: There is no sanitization of the data retrieved from the web before it is passed back to the agent.
Audit Metadata