troubleshooting
Warn
Audited by Gen Agent Trust Hub on Oct 8, 2026
Risk Level: MEDIUMDATA_EXFILTRATIONEXTERNAL_DOWNLOADSDYNAMIC_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [DATA_EXFILTRATION]: The skill searches for and reads several configuration files in the user's workspace, including .claude/settings.json, .vscode/launch.json, and .gemini/settings.json. These files are used to store application and environment settings and may contain sensitive data or custom configurations.
- [EXTERNAL_DOWNLOADS]: The skill fetches documentation and troubleshooting guides from the official Chrome DevTools GitHub repository (https://github.com/ChromeDevTools/chrome-devtools-mcp/blob/main/docs/troubleshooting.md).
- [DYNAMIC_EXECUTION]: The skill recommends using npx to download and execute the chrome-devtools-mcp package from the NPM registry to perform diagnostics. It also uses the gh command-line tool to interact with GitHub issues.
- [INDIRECT_PROMPT_INJECTION]: The skill provides a surface for indirect prompt injection by processing external documentation and user-provided configuration files.
- Ingestion points: troubleshooting.md (via GitHub URL) and local files: .mcp.json, gemini-extension.json, .claude/settings.json, .vscode/launch.json, .gemini/settings.json.
- Boundary markers: None present; the agent is instructed to read and interpret the file contents directly.
- Capability inventory: Execution of shell commands via npx and gh, and browser control via MCP tools (list_pages, etc.) as described in SKILL.md.
- Sanitization: The skill does not specify any sanitization, filtering, or validation steps for the content ingested from external sources or local files.
Audit Metadata