按任务自动调参 Qwen3.8 reasoning-budget 的守门人
-
Updated
Nov 19, 2026 - Python
按任务自动调参 Qwen3.8 reasoning-budget 的守门人
把 GLM-5.2 接进任意 OpenAI 格式编码 agent,tool-call 不再静默错配——一层薄协议适配,harness 的 OpenAI 代码路径一行不改。
为 coding agent 改动评审做双审分歧分流的开源命令行工具,只有模型意见相反的改动才需要人看
v0.2.0: detected model id now sent in probe payloads; malformed endpoint responses classify cleanly (no raw tracebacks); clean exit-2 usage errors for bad --out; new no_tool probe for over-eager tool-call detection (suite cn-tc-v2).
v0.4.0:截断落在逗号、冒号或键中间时只丢弃残缺成员、保留完整字段,可修复的流不再以退出码 1 失败;字符串值里的代码围栏不再被误当成外层 fence 而改写整段参数;上游压缩的 SSE 会先解压再修复,Accept-Encoding 不再让修复静默失效。
一条命令对比 DeepSeek/Qwen/Kimi/GLM 对你输入的分词数、上下文适配率与含缓存折扣的成本。
Run-level exposure gauge for coding agents on local models — tracks wall-clock, state mutations, and token burn since last human review, forces a pause at thresholds.
为 coding agent 标出每周影响配置的国产模型 drop 变更
国产大模型工具调用契约差分测试器
Capability-gap diagnostician for locally-served CN open-weights models (DeepSeek/Qwen3/Kimi K3/GLM) — attributes your local model's felt 'dumbness' to a specific misconfiguration layer so you fix one layer, not the whole stack.
ctxfeed is a local MCP project-context backend that shards a whole repo into GLM-5.2's 1M-token window with cache-aware ingest ordering
To associate your repository with the cn-model-devtools topic, visit your repo's landing page and select "manage topics."