mk:chrome-profile
Connect browser automation to a named local Google Chrome profile through CDP without guessing or replacing the user's session.
| Source | .claude/skills/chrome-profile/SKILL.md |
| Owner | utility |
| Runtime | claude-code |
| Risk | medium |
| Depends on | mk:agent-browser |
What This Skill Does
Targets a named local Chrome profile for browser automation through the Chrome DevTools Protocol.
It preserves that profile's cookies, extensions, and existing logged-in sessions while connecting
mk:agent-browser to a dedicated debug port.
When to Use
- Automate a site with the user's selected Work or Personal Chrome profile.
- Use a saved browser session rather than a clean automation profile.
Do not use it for clean headless browsing; use mk:agent-browser directly instead.
Safety Contract
The skill lists available profiles from Chrome's Local State data and proceeds only when the requested name has exactly one match. It never kills an existing Chrome process and stops on a debug-port conflict rather than guessing.
Workflow
- Discover and select one profile.
- Start Chrome with
--remote-debugging-portwhen needed. - Run
agent-browser connect <port>. - Continue normal browser automation with
mk:agent-browser.
See the canonical mk:chrome-profile skill for platform commands and recovery details.
mk:agent-browser
AI-agent browser automation via agent-browser (Chrome/CDP). Accessibility-tree snapshots with @eN refs, sessions, auth vault, MCP profiles, React/vitals, cloud providers, and specialized workflows for dogfooding, Electron, and Slack.
mk:playwright-cli
Session-persistent browser automation via Playwright CLI — form filling, screenshots, data extraction, multi-step flows.