Automated README summary
Automatically summarized from the README on 28 Oct 2026; not a run by jevbooks. Run instructions and author-reported outcomes still need verification.
What it is
A DeepSeek Harness plugin for coding agents that judges each goal round at the turn-stopping seam, so the goal loop stops on measured evidence rather than a round counter.
How it uses Jev
At agent/turn-stopping, when a goal is active and the closing turn is round N≥1, one Jev fan-out asks five fixed questions over a bounded per-round digest: Nouls on objective achieved and evidence in transcript, a progress Score, a same-obstacle Noul, and an action Choice. Noul thresholds route the verdict; the Choice only breaks ties. In shadow mode the verdict is logged to telemetry; in enforce mode a fixed-template steer note (verify-complete or report-blocked) is submitted to the next-step boundary.
Primitives:noulscorechoice
Technique worth stealing
Keep instruction strings constant and put transcript text only in the state data field, so tool output cannot rewrite the questions; deterministic threshold routing replays exactly over logged answers.
Try it
dsh plugin --profile add dsh-jev-goal-judge, then set mode: shadow in cordis.patch.yml and read /jev-goal-stats.
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 material0.89
- Shows a System One patternFound in material0.94
- Describes handling uncertaintyFound in material0.97
- Reports measured numbersFound in material0.94
- Is a runnable toolFound in material0.97
- Top-10 learning shortlist signalunclear0.56
- Is a replica of the modelnot found in the excerpt0.11
- Solves a common problem (0–2)score on a 0–2 scale1.04 / 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 28 Oct 2026.