From radar
Build and maintain a catalogue of AI tools, features, and techniques from external sources. Runs independently of your personal context.
How this skill is triggered — by the user, by Claude, or both
Slash command
/radar:radar-scanThe summary Claude sees in its skill listing — used to decide when to auto-load this skill
Build and maintain a catalogue of AI tools, features, and techniques from external sources. Runs independently of your personal context.
Build and maintain a catalogue of AI tools, features, and techniques from external sources. Runs independently of your personal context.
$ARGUMENTS — Optional:
--sources <all|feeds|manual> — Which sources to scan (default: all)--days N — How far back to look in time-based sources (default: 7)Parse from $ARGUMENTS if provided. Default to --sources all --days 7.
Check if the ai-brain MCP tools are available (try calling get_lists).
If brain MCP is available:
Call get_lists and look for lists with names starting with [Radar]. Create any that are missing:
[Radar] Inbox — raw links dropped by user for enrichment[Radar] Claude Code — features, settings, tips[Radar] MCP Ecosystem — servers, plugins, integrations[Radar] AI Tools & Techniques — broader tools, prompting, workflowsUse create_list for each missing list.
If brain MCP is unavailable:
Use local JSON file at ~/.claude/radar-catalogue.json. Read it if it exists, or initialize with empty structure.
If ~/.claude/radar-catalogue.json does not exist but ~/.claude/scout-catalogue.json does, rename ~/.claude/scout-catalogue.json to ~/.claude/radar-catalogue.json and use it.
{
"lists": {
"[Radar] Inbox": { "items": [] },
"[Radar] Claude Code": { "items": [] },
"[Radar] MCP Ecosystem": { "items": [] },
"[Radar] AI Tools & Techniques": { "items": [] }
},
"lastUpdated": null
}
Load all items from [Radar] lists to build a set of known URLs for deduplication.
Brain mode: Call get_list for each [Radar] list. Collect all item URLs into a set.
Local mode: Read from the JSON file.
Run node "${CLAUDE_PLUGIN_ROOT}/bin/workflow-analyzer/dist/cli.js" scan-deps --since ${DAYS} --output /tmp/workflow-analyzer-deps.json. Read the output JSON. If the bundled binary is not available, fall back to npx @flippyhead/workflow-analyzer@latest scan-deps --since ${DAYS} --output /tmp/workflow-analyzer-deps.json.
If the command fails or is not available, log a warning and skip to Step 3 — dependency scanning is additive, not required.
If GITHUB_TOKEN is not set in the environment and the scan-deps output shows rateLimited > 0, print: "GitHub API rate limited — scanned [reposResolved] of [packageCount] dependencies. Set GITHUB_TOKEN for full scanning."
For each entry in the releases array:
release.body (release notes) and repoDescription to assess relevancesource: "dependency-changelog" and an additional relevanceHint of "direct dependency"release.url as the item URL for deduplication against existing catalogueSkip this step if --sources manual was specified.
Limit to 10-15 items per source. If a source fails (timeout, rate limit, format change), log a warning and continue to the next source — never fail the entire run.
When a source fails, print a clear one-line message: "Source [name] unavailable: [reason]. Continuing with remaining sources."
Anthropic changelog/blog:
Use WebFetch on https://docs.anthropic.com/en/docs/about-claude/models and https://www.anthropic.com/news to find recent releases and feature announcements. Extract title, URL, and a one-line description for each.
Hacker News:
Use WebFetch on the Algolia API:
https://hn.algolia.com/api/v1/search_by_date?query=claude+code&tags=story&numericFilters=created_at_i>${DAYS_AGO_TIMESTAMP}https://hn.algolia.com/api/v1/search_by_date?query=anthropic+mcp&tags=story&numericFilters=created_at_i>${DAYS_AGO_TIMESTAMP}https://hn.algolia.com/api/v1/search_by_date?query=ai+agent+tool&tags=story&numericFilters=created_at_i>${DAYS_AGO_TIMESTAMP}Extract title, URL (use url field, fall back to HN comment URL), and points as a quality signal. Only keep items with 5+ points.
GitHub:
Use WebSearch for:
Extract repo name, URL, and description.
YouTube:
Use WebSearch for:
Extract video title, URL, and channel name.
For each result across all sources: skip if the URL already exists in the catalogue (deduplication).
Also deduplicate across sources within this run — if the same URL was found by both HN and GitHub in this run, only catalogue it once.
Skip this step if --sources feeds was specified.
Brain mode: Call get_list for the [Radar] Inbox list. For each item with status "open":
WebFetch to get the page contentdescription[Radar] category list using create_list_item with the enriched fieldsupdate_list_itemLocal mode: Process items in the Inbox array, move to the appropriate category array.
For each new catalogue entry (from Step 3 or Step 4), set the properties field:
{
"category": "<one of: claude-code, mcp, api, agent-sdk, prompting, tooling, workflow, general-ai>",
"relevanceHints": ["<free-text tags describing what workflows/goals this helps with>"],
"source": "<one of: anthropic-changelog, hackernews, github, youtube, manual, dependency-changelog>",
"discoveredAt": "<ISO date>"
}
Choose category based on the content:
claude-code — Claude Code features, settings, shortcuts, pluginsmcp — MCP servers, protocols, integrationsapi — Claude API features, SDK updatesagent-sdk — Agent building tools and frameworksprompting — Prompting techniques, system prompts, skill designtooling — Developer tools, CLI utilities, browser extensionsworkflow — Workflow patterns, automation techniques, productivity methodsgeneral-ai — Broader AI developments, models, researchChoose relevanceHints based on what kinds of work this would help with (e.g., "browser automation", "code review", "testing", "deployment", "project management").
Brain mode: Use create_list_item with url, description, and properties set.
Local mode: Append to the appropriate list in the JSON file. Write the file when done.
Output a brief summary:
[Radar] listsnpx claudepluginhub flippyhead/radar --plugin radarGuides creation, editing, and verification of skills for AI coding agents using test-driven development with subagent scenarios. Use when authoring or debugging skills.