Competitor Launch Radar

What it does

Competitor Launch Radar performs a weekly scan of Product Hunt RSS and Hacker News (Algolia) to surface brand-new AI agent framework entrants that fall outside a predefined 9-framework cohort. It filters noise, applies a 10-upvote/points noise floor, deduplicates against an LRU announced state, classifies entrants (framework / mcp / product), and writes a digest article plus optional notification.

When to use it

Run weekly (operator schedule: Monday) to catch launches the core watchlist would miss — useful for product & competitive intelligence teams tracking emergent frameworks, MCP tools, and agent-powered products. Use dry-run mode to preview state and article output without sending notifications.

What's included

• Scripts: lightweight fetch + parse routines and state handling; fallback WebFetch modes for environments that block direct curl.
• References: methodology and state schema documented in-skill; sample article templates and logs.
• Instructions: full end-to-end steps: fetch → filter (suppression, keywords, noise floor) → classify → sort → write article → update LRU state → notify & log.

Compatible agents

Works with monitoring/automation agents that can run scheduled scripts and write to local memory — suitable for CLI-based automation, GitHub Action-style cron jobs, or autonomous agents with HTTP read capability.