| name | configure |
| description | Configure pattern-radar sources, weights, and domains. |
| allowed-tools | mcp__pattern-radar__configure_sources |
Configure Your Radar
Customize which sources to scan and how to score relevance.
Setting Domains
Your domains determine what's relevant to you:
configure_sources(domains: ["TypeScript", "React", "AI", "distributed systems"])
Domains can be:
- Programming languages (TypeScript, Python, Rust)
- Frameworks (React, Next.js, FastAPI)
- Technologies (WebAssembly, GraphQL, Kubernetes)
- Concepts (distributed systems, machine learning, security)
Source Configuration
Enable/disable sources:
configure_sources(hackernews_enabled: true, github_enabled: true)
Adjust weights (0-2, higher = more influence):
configure_sources(
hackernews_weight: 1.5, # Prioritize HN
github_weight: 0.5 # De-prioritize GitHub
)
GitHub language filter:
configure_sources(github_languages: ["TypeScript", "Rust", "Python"])
Profile Integration
If shared-memory is installed, pattern-radar reads your profile automatically:
- Languages from
technical.languages - Frameworks from
technical.frameworks - Domains from
knowledge.domains - Interests from
knowledge.interests
You can override or supplement with configure_sources.
GitHub Token
For higher rate limits (5,000/hour vs 60/hour):
- Create token at https://github.com/settings/tokens
- Run pattern-radar setup:
node run.js - Enter token when prompted
Token is saved to ~/.config/brain-jar/pattern-radar.json
Current Config
View current configuration by calling any radar tool—it shows your active settings.
Reset to Defaults
Delete the config file:
rm ~/.config/brain-jar/pattern-radar.json