radar-onboard
Skill4 starsChanged 5 months agoMITView on GitHub
Run this FIRST, before any other radar skill. Sets up Competitive Intelligence Radar by ASKING whether BrightData and Notion are connected, live-testing one real call to each, detecting which MCP prefix is actually live, and writing radar.config.json with the user's business context. Use when the user runs /radar-setup, says "set up the radar", "onboard", or when any radar skill reports that radar.config.json is missing. Does not assume tools are connected; it verifies.
In this repository8 other files it gives its agents.
- CLAUDE.mdCLAUDE.md
- radar-diff-changesSkill
- radar-merge-briefingSkill
- radar-plan-targetsSkill
- radar-provision-notionSkill
- radar-publish-councilSkill
- radar-publish-notionSkill
- radar-render-dashboardSkill
The file
We split this skill into the rules it gives an agent
6 rules. About 0 are common, 3 are distinctive, and 3 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈3
Distinctive · estimate
This author's own choices. Worth a look.
3
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 9
Never assume a connection.
- Distinctive: line 20
ASK about connections, do not assume: "Have you connected BrightData and Notion?
- Very specific: line 29
Use query like {"query": "site:example.com", "engine": "google"}.
- Distinctive: line 39
If EITHER test fails, STOP.
- Very specific: line 55
Ask where in Notion to put the databases: a parent page id or "create at workspace level." Store as notion.parent_page_id (may be empty for …
- Very specific: line 59
Write radar.config.json at the project root with everything gathered and the two verified prefixes.
≈ “Common” is an estimate: at least 1 in 5 of the 372,429 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
radar-onboard
Automation level: CHECKPOINT. You ask the user real questions and you stop if a live MCP test fails. Never assume a connection. See automation-levels.md.
Goal: produce a complete radar.config.json (schema in
config-schema.md) with verified MCP prefixes, so every later
skill just works.
Steps
- Greet and explain in two lines what the radar does and that it needs BrightData (web data) and Notion (publishing). Keep branding generic; this plugin serves any business.
- ASK about connections, do not assume: "Have you connected BrightData and Notion? I usually see two paths: through this plugin, or through the Claude Desktop app connectors. I will test both."
- Detect the live BrightData prefix. Try a tiny real call in this order until one works, using a
harmless query:
- Plugin path:
mcp__plugin_competitive-intelligence-radar_brightdata__search_engine - Desktop path:
mcp__claude_ai_Bright-Data__search_engine - Any other connected BrightData server visible in this session (a UUID prefix is normal).
Use query like
{"query": "site:example.com", "engine": "google"}. The FIRST prefix that returns a result is the live one. Recordmcp.brightdata_prefixandmcp.brightdata_verified_at.
- Plugin path:
- Detect the live Notion prefix. Try a tiny real read, for example
notion-get-usersor anotion-searchwith a simple query, in this order:mcp__plugin_competitive-intelligence-radar_notion__mcp__claude_ai_Notion__- Any other connected Notion server (UUID prefix).
Record
mcp.notion_prefixandmcp.notion_verified_at.
- If EITHER test fails, STOP. Do not write a half-config that pretends a connection exists. Tell the
user exactly how to connect:
- BrightData: get a free token at brightdata.com, then either set
BRIGHTDATA_API_TOKENin the environment so this plugin's.mcp.jsonserver starts, or add BrightData in the Claude Desktop app connectors. Re-run/radar-setupafter. - Notion: connect Notion via OAuth, either through this plugin's Notion server or the Desktop app connectors. Re-run after.
- BrightData: get a free token at brightdata.com, then either set
- Collect business context interactively and confirm it back:
- Business name, domain, one-line description, 2-letter geo.
- Tracked keywords (the searches where ranking matters).
- Competitor list: name + domain. Remind the user of the cap of 8 (cost-control.md); if they give more, keep the first 8 and say so.
- Scout depth: ask whether to include social sentiment, financials, hiring. Default all to off for a cheap lean run; the user can turn them on.
- Ask where in Notion to put the databases: a parent page id or "create at workspace level."
Store as
notion.parent_page_id(may be empty for workspace level). Do NOT create databases here; that isradar-provision-notion's job after the user approves the schema. - Write
radar.config.jsonat the project root with everything gathered and the two verified prefixes. Confirm the file path and summarize what was saved (no secrets). - Tell the user the next step: run
/radar-setupcontinues intoradar-provision-notionto review and approve the Notion schema, then/radarfor the first gather.
Edge cases
- Both prefixes live but pointing at different connection methods (for example BrightData via plugin, Notion via desktop): that is fine. Record each independently.
- A prefix returns an auth error rather than a clean result: treat as a failed test, stop, and give the connect instructions for that service.
- Re-running onboarding: load the existing config if present, prefill answers, and let the user edit rather than starting blank. Overwrite only after confirmation.
- User refuses a deep source: leave its
scout_depthflag false; never enable a source the user did not ask for.
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

