Viger1/dsh-review

README excerpt

dsh-review Packaged multi-agent adversarial code review for DeepSeek Harness. English 中文 Status: M0 — the review tool works end to end. The problem dsh ships the primitives for multi-agent work — subagents, workflows, ralph loops — but nothing that packages them into a review you can trust. Run naive parallel reviewers over a diff and you get a wall of plausible-sounding findings, most of them wrong; triaging them costs more than the review saved. "AI code review produces too many false positives" is the reason people stop using it. The method Two stages, and the second is the one that matters: 1. Find — several reviewers in parallel, each with a distinct lens (correctness, lifecycle/concurrency, API-contract conformance, security), each reporting findings with a concrete failure scenario rather than a style opinion. 2. Verify adversarially — every finding gets its own verifier whose job is to refute it: read the real code, reproduce if possible, and default to "not real" when the evidence is ambiguous. Only survivors are reported. Measured on this repo's two sibling plugins (dsh-preview, dsh-pilot): 73 agents, 49 confirmed findings, 14 refuted. Two of the confirmed ones were only …

Category