Automated README summary
Automatically summarized from the README on 29 Oct 2026; not a run by jevbooks. Run instructions and author-reported outcomes still need verification.
What it is
An OpenCode V2 server plugin that trims MCP server, MCP tool, and skill descriptions from each outgoing model request, keeping only candidates Jev judges relevant to the current user prompt. Aimed at users running many MCP servers who want fewer input tokens and fewer wrong-tool calls.
How it uses Jev
Before each model call, the plugin sends the user prompt as Jev state with batched `noul` (yes/no) questions: phase 1 asks whether each MCP server is needed, phase 2 asks about individual tools and skills. Candidates scoring below the configured threshold are removed from `event.system` (Code Mode catalog, mcp instructions, skills) and deleted from `event.tools`.
Primitives:noul
Technique worth stealing
Two-phase gating: cheaply drop whole MCP servers first, then score only surviving servers' tools and skills, with fail-open on any error.
Try it
Export JEV_API_KEY and run `opencode --standalone`, or use `npm run fake-jev` with JEV_URL for offline testing.
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 projectFound in material0.76
- Shows a System One patternunclear0.42
- Describes handling uncertaintyFound in material0.66
- Reports measured numbersnot found in the excerpt0.02
- Is a runnable toolFound in material0.95
- Top-10 learning shortlist signalnot found in the excerpt0.27
- Is a replica of the modelnot found in the excerpt0.03
- Solves a common problem (0–2)score on a 0–2 scale1.19 / 2
- About JevFound in material0.98
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 29 Oct 2026.