configure
Configure pattern-radar sources, weights, and domains.
When & Why to Use This Skill
This Claude skill enables the precise configuration of the Pattern-Radar tool, allowing users to customize data sources, relevance weights, and technical domains. It optimizes how the agent monitors platforms like GitHub and Hacker News, ensuring that the insights generated are perfectly aligned with specific programming languages, frameworks, and technological interests.
Use Cases
- Personalizing technical trend monitoring by prioritizing specific sources like Hacker News or GitHub based on information quality.
- Filtering repository scans to focus exclusively on relevant programming languages such as TypeScript, Rust, or Python.
- Defining custom interest domains like 'Distributed Systems' or 'AI' to ensure the agent captures the most relevant industry developments.
- Managing GitHub API authentication to enable higher rate limits for intensive technical research and data gathering.
- Synchronizing the agent's research focus with a user's professional profile for automated, context-aware knowledge discovery.
| name | configure |
|---|---|
| description | "Configure pattern-radar sources, weights, and domains." |
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