Automated README summary
Automatically summarized from the README on 31 Oct 2026; not a run by jevbooks. Run instructions and author-reported outcomes still need verification.
What it is
A small, zero-dependency Node.js demo for a talk on decision models. It runs four coding-agent scenarios (issue triage, tool-call gating, fix ranking, task routing) against TypeSafe Jev, Microsoft-Decision-1, the OpenAI Decisions API, or a mock, and compares them with a generative LLM baseline.
How it uses Jev
Jev is one of several providers behind a single adapter layer, selected via DECISION_PROVIDER=typesafe. Scenario code sends state (issue text, tool call, candidate fixes, task) plus typed questions and reads back per-option probabilities; thresholds in code decide paging, allow/ask/block, ranking, and routing. The README does not detail Jev-specific request handling beyond the endpoint and model name.
Primitives:choicescorenoul
Technique worth stealing
One adapter layer over multiple decision-model vendors so scenario code stays unchanged; typed questions return probabilities you gate on in code.
Try it
git clone the repo, then run npm run demo, npm run compare, npm test (Node.js 20+, no keys).
Model judgments from README excerpts
Project signals use repository metadata, an existing summary and the first 700 README characters, not the full project. “Not found” means not detected in that material; it does not establish absence. These are the original judgments with recorded models and dates, not source audits or run results.
Models & evaluation dates
- Inclusion gate
jev-1.14.0- Project signals
jev-1.14.0- Category
jev-1.14.0- Pattern tags
jev-1.14.0
Evidence
These signals use repository metadata, an existing summary and at most 700 cleaned README characters. ≥ 0.60 means reported in that excerpt; ≤ 0.40 means not found there. The middle band is inconclusive. None is a source-code audit or a test result. The runnable-tool question excludes research experiments; the recommendation question asks for a learning shortlist, not a quality verdict.
- Jev is core to the projectnot found in the excerpt0.06
- Shows a System One patternFound in material0.61
- Describes handling uncertaintynot found in the excerpt0.03
- Reports measured numbersnot found in the excerpt0.03
- Is a runnable toolnot found in the excerpt0.26
- Top-10 learning shortlist signalnot found in the excerpt0.25
- Is a replica of the modelnot found in the excerpt0.03
- Solves a common problem (0–2)score on a 0–2 scale0.42 / 2
- About JevFound in material0.97
no-judgment band (0.40–0.60)
no-judgment band (0.40–0.60)
The Jev model and date for each judgment stage are listed above. Card written by DeepSeek V4.1 Flash from the README on 31 Oct 2026.