dev-tools-skill
Pass
Audited by Gen Agent Trust Hub on Jun 23, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill consists of markdown documentation and guidance for using Chrome DevTools MCP tools to perform web application debugging. All analyzed files contain only instructional content.
- [COMMAND_EXECUTION]: The instructions leverage browser automation tools such as "evaluate_script", "click", and "fill" to interact with and inspect pages. These tools are scoped to the browser environment, and the instructions explicitly advise the agent to explain actions in plain language and seek user confirmation before performing destructive actions.
- [PROMPT_INJECTION]: The skill contains a potential surface for indirect prompt injection as it processes data from external web pages (DOM snapshots, console logs, and network payloads). This is an inherent property of browser debugging tools and is mitigated by the skill's guided workflow and safety rules.
- Ingestion points: "take_snapshot", "list_console_messages", and "get_network_request" mentioned in "SKILL.md" and "references/mcp-cheatsheet.md".
- Boundary markers: Not explicitly defined for delimiting untrusted web content.
- Capability inventory: Browser interaction and scripting tools including "evaluate_script", "click", "fill", and "press_key".
- Sanitization: No explicit sanitization or filtering of external web content is described in the instructions.
Audit Metadata