Dev builds JCR on decision model Jev so agents skip skill files and query a capability tree directly
gaganghotra_ · x · 2026-09-21
What JCR is
Niaz Morshed built JCR (Jev Capability Resolver) on top of TypeSafe's decision model Jev (early access since Sep 15, makes structured decisions with probabilities code can act on): a recursive loop searches a capability tree and brings back only task-relevant context to the agent.
The problem it solves
- With the usual skills mechanism, an agent doing three operations (upload a Cloudflare Worker, create a Stripe Checkout Session, send a Resend email) reads six documents — three SKILL.md files plus reference guides — loading setup notes, alternatives, and irrelevant examples into context.
- JCR moves that lookup out of the agent's working context, complementing rather than replacing skills.
More from coding & agent
- Typed Evals: open-source framework for calibrated Jev-powered LLM evaluation — Charming_Group_2950 · 2026-09-21
- diagram-design skill for Claude Code hits 41.5k GitHub stars with editorial-quality diagrams — bibryam · 2026-09-21
- Qwen-Image 2.1 open-weights 7B model gets Day-0 ComfyUI support, runs in 30s on a laptop — Lexius2129 · 2026-09-21
- Dev builds bare-metal autonomous agent system after OpenAI refused to refund lost quota — epicskyes · 2026-09-21
- Open-source 'jev' tool automates computer use with no LLM, sub-second latency — _AustinCalvert_ · 2026-09-21
- Ex-Stripe Exec: Ready Your Business for Agent Buyers With Checkout, Payment Element and Invoices — jeff_weinstein · 2026-09-21