Automated README summary
Automatically summarized from the README on 26 Sept 2026; not a run by jevbooks. Run instructions and author-reported outcomes still need verification.
What it is
unwedge is a circuit breaker for AI coding agents. It watches Claude Code, Codex CLI or custom agent loops for stuck behaviour (repeated failing commands, ignored errors, no progress) and alerts before the loop burns budget.
How it uses Jev
On suspicious turns, unwedge optionally asks jev typed questions about the session state: is the agent still getting closer, is the work already done, is it drifting. jev returns probabilities only; unwedge uses them in its policy to decide hints or stops.
Primitives:choicescorenoul
Technique worth stealing
Combine code-computed loop signals with typed model judgments on suspicious turns only, using hysteresis and hint-before-stop.
Try it
uv tool install unwedge; unwedge scan; or install the Claude Code plugin.
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-1.13.0- Project signals
jev-1.13.0- Category
jev-1.13.0- Pattern tags
jev-1.13.0
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 excerpt0.04
- Shows a System One patternnot found in the excerpt0.28
- Describes handling uncertaintynot found in the excerpt0.14
- Reports measured numbersnot found in the excerpt0.09
- Is a runnable toolFound in material0.86
- Top-10 learning shortlist signalnot found in the excerpt0.15
- Is a replica of the modelnot found in the excerpt0.07
- Solves a common problem (0–2)score on a 0–2 scale1.43 / 2
- About JevFound in material0.88
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 26 Sept 2026.