{"manifest":{"manifestVersion":1,"name":"runchr/competitive-research-agent","revision":1,"version":"0.1.0","tier":"draft","tools":[{"type":"mcp-server","ref":"browserbase-mcp","install":"{\n  \"mcpServers\": {\n    \"browserbase\": {\n      \"command\": \"npx\",\n      \"args\": [\"-y\", \"@browserbasehq/mcp-server-browserbase\"],\n      \"env\": {\n        \"BROWSERBASE_API_KEY\": \"<api-key>\",\n        \"BROWSERBASE_PROJECT_ID\": \"<project-id>\"\n      }\n    }\n  }\n}"},{"type":"skill","ref":"graphify"},{"type":"framework","ref":"scrapling"}],"model":"claude-sonnet-4.5","harness":"claude-code","systemPrompt":"You are a competitive research agent. You read pages a browser can reach, extract only the fields you were asked for, and record where each value came from.","instructions":"For each competitor, open the pricing page with the browser, extract the plan names and prices with the scraping framework, and append a row to the notes file with the source URL. Never invent a price; leave the cell empty and flag it."}}