- Главная
- Плагины
- Разработка и инструменты плагинов
- dsh-resource-guard
dsh-resource-guard
wth373737/dsh-resource-guard
Pre-flight resource gate for tool calls: estimates peak memory and VRAM from the command line, the scripts it references and per-binary costs, compares that with the machine live headroom and allows, asks or refuses; a budget block in the system prompt lets the model size concurrency, a runtime breaker warns when a call crosses the line, and measured peaks feed back into later estimates. Windows only. Its HTTP routes have no DSH access check, so a web page open in your browser could call them.
Установка
dsh plugin --profile web add github:wth373737/dsh-resource-guardREADME
resource-guard
A resource gate for DeepSeek Harness: it estimates a tool call's memory and VRAM cost before the call runs, asks for confirmation when the call would be heavy, and feeds a live hardware budget into every session so the model plans concurrency against the real machine instead of guessing.
English
Why this exists
On 2026-11-09 a single command — py -3.12 align_parallel.py — hid max_workers=6 inside the file. Six qwen3-aligner processes at roughly 6 GB of commit each, on top of a running game and a video editor, exhausted a 66.9 GB commit limit in eight seconds. DWM crashed, Explorer hung, the desktop froze, and the session died mid-turn.
The command line looked harmless. The damage was invisible until the machine was already gone.
What it does
| Mechanism | Hook | Behaviour |
|---|---|---|
| Pre-flight gate | tools/pre-execute | Estimate peak → compare with current headroom → allow / ask / deny |
| Runtime breaker | 1 s sampling + tools/post-execute | On a breach, append a warning to the tool result and tell the model to converge concurrency |
| Budget injection | systemPrompt.context | Per-step memory / VRAM / safe-concurrency block |
| Profile learning | local JSON | Records the measured increment per signature; estimates improve with use |
How the estimate is built
- Explicit signals on the command line —
max_workers=6,-j 8,xargs -P,-ThrottleLimit. - The contents of referenced scripts — when the command names a
*.py|ps1|sh|js|bat|cmdfile that exists and is ≤ 1 MB, the first 64 KB are scanned for worker counts, binary names and torch/onnx markers. This is the layer built specifically for the incident above: the concurrency was not on the command line. - Per-binary unit costs —
qwen3-aligner6 GB/process (measured on this machine), Python + torch 3.2 GB,ffmpeg1.2 GB,ollamascaled by model size. - History — once a signature has run, the measured increment corrects the rule estimate.
Installation
Two ways in. The plugin market and the install command use the first; the second is for working on the code.
From GitHub (recommended)
In DSH, open Plugins → + Add plugin and enter:
github:wth373737/dsh-resource-guard
Then restart DeepSeek Harness once — a replaced package only loads a fresh module generation on restart.
From a local checkout (development)
A link: install points the profile at the directory itself, so edits are live after a restart and git pull is enough to update. Keep it somewhere stable — not a temporary folder or a session workspace.
- Clone it:
git clone https://github.com/wth373737/dsh-resource-guard.git - Install the bundle into the running profile by absolute directory path — through the plugin manager's
install_bundle, or by asking the agent to do it with that path. - Restart DeepSeek Harness once.
Either way, confirm it is live: the next request's system context carries a resource-budget block (currently Chinese) with commit, VRAM and a suggested concurrency.
Configuration
Set these under config in the bundle's cordis.patch.yml; the plugin manager reloads them.
| Key | Default | Meaning |
|---|---|---|
enabled | true | Master switch |
mode | ask | ask to prompt, deny to refuse outright, warn to record only |
scope | heavy | heavy gates known-expensive tools, all gates every call |
reserveGb | 6 | Headroom reserved when judging physical pressure |
askRatio | 0.8 | Projected commit / limit that triggers a question |
denyRatio | 0.94 | Projected commit / limit that triggers a refusal |
vramAskRatio | 0.85 | VRAM question threshold |
vramDenyRatio | 0.97 | VRAM refusal threshold |
minEstimateGb | 4 | Estimates below this are never gated |
perWorkerGb | 3.2 | Assumed peak commit of one worker when sizing safe concurrency; replaced by the measured median once history holds ≥ 3 samples |
breaker | true | Runtime monitoring |
breakerRatio | 0.92 | Observed commit ratio that counts as a breach |
injectBudget | true | Inject the live budget block |
learn | true | Feed measured peaks back into the estimator |
autoTune | suggest | off / suggest (write the rationale document only) / apply (also write the recommended values into this config) |
tools | [] | Extra tool names always treated as heavy |
debug | false | Also log to stderr |
Settings page
Installing the bundle adds Settings → Resource Guard.
That page is contributed by the bundle's client half (dsh.client + client.js), not generated from the Host config: everything under Settings is registered by client plugins into settings.section, so a Host-only plugin gets no page at all — its configuration exists but nothing can show or change it. v0.3.0 adds the client half for that reason.
The page shows the live budget (commit, VRAM, safe concurrency), the hardware-derived suggestion with a fill recommended values action, and every threshold as an editable field with its explanation. It reads and writes through a fenced same-origin route (/resource-guard/api, requiring the x-resource-guard-ui header). The Host side reads the live values and applies changes through ctx.configEditor — the same Loader path the plugin manager uses — so a save lands in the profile patch exactly like a hand edit. The settings service is kept as a fallback only.
After installing a new version, refresh the page once — the client bundle is loaded at boot.
The page follows the system language (navigator.language): a Chinese system renders the Chinese page, anything else English. It is decided when the page loads, so changing the system language needs a reload.
Recommendation document
At startup the plugin measures the machine and writes $DSH_HOME/resource-guard/recommendation.md: the recommended thresholds plus, for every one, the hardware fact it came from. For example:
askRatio→ 0.78 — commit limit 66.9 GB ÷ 62.9 GB physical = 1.06, page file only 4 GB, so there is almost no overflow room and the ask line drops to 78%.
autoTune: suggest (the default) only writes that document. apply also writes the values into this config once per hardware fingerprint, recorded in applied.json; off skips the whole step.
State
$DSH_HOME/resource-guard/
history.json— for each signature, the typical cost (a running average, used for estimates) and the peak seen (used to warn). Safe to delete; losing it only falls back to rule-only estimates.recommendation.md— the hardware-derived suggestions for this machine, regenerated at each start.applied.json— the hardware fingerprint of the lastautoTune: applyrun.guard.log— latest run log, rotated at 512 KB.
Known limits
- Opaque commands are still guessed. If concurrency is computed at runtime (read from a config, derived from CPU count), static analysis cannot see it. The runtime breaker is the backstop.
- VRAM is sampled slowly. One
Get-Countermeasurement blocks for 1–1.5 s, so VRAM refreshes every 10 s and memory every 3 s. Short spikes can be missed. - No hard kill. v1 warns and asks the model to converge. Safely killing only the processes a call spawned needs process-tree ownership; killing the wrong PID is worse than the incident. Deferred to v2.
- Schema loading is best-effort. The settings page needs the Harness's own
schemastery; when it cannot be resolved the plugin still runs, just without the form. - Two persistent PowerShell samplers, roughly 40–60 MB each.
- The injected budget block is Chinese only. The prompt context exposes no UI locale, so the block is not localized yet.
Requirements
DeepSeek Harness with the plugin manager; Windows (the samplers use GlobalMemoryStatusEx and WDDM performance counters).
Uninstall
Remove the dsh-plugin-resource-guard bundle in the plugin manager, then delete $DSH_HOME/resource-guard/.
License
MIT © 2026 wth373737 — see LICENSE.
中文
为什么有这个东西
2026-11-09,一条 py -3.12 align_parallel.py 的命令,把 max_workers=6 藏在了文件里。六个 qwen3-aligner 进程、每个约 6 GB 提交内存,加上当时正在跑的游戏和剪辑软件,八秒内击穿了 66.9 GB 的提交上限。DWM 崩溃、资源管理器挂起、桌面冻住,会话在中途死掉。
命令行看上去人畜无害。等发现的时候,机器已经没了。
它做什么
| 机制 | 挂点 | 行为 |
|---|---|---|
| 事前门禁 | tools/pre-execute | 估算峰值 → 与当前余量比较 → allow / ask / deny |
| 运行时熔断 | 1 秒采样 + tools/post-execute | 越线时在工具结果里追加告警,并要求模型收敛并发 |
| 预算注入 | systemPrompt.context | 每步可见的内存 / 显存 / 安全并发 |
| 画像学习 | 本地 JSON | 按签名记录实测增量,越用越准 |
估算怎么来
- 命令行的显式信号 ——
max_workers=6、-j 8、xargs -P、-ThrottleLimit。 - 读被引用脚本的内容 —— 命令行里出现存在且 ≤ 1 MB 的
*.py|ps1|sh|js|bat|cmd文件时,读前 64 KB 继续找并发数、二进制名和 torch/onnx 标记。这一层就是为上面那次事故加的:并发数根本不在命令行里。 - 按二进制给单价 ——
qwen3-aligner6 GB/进程(本机实测)、Python + torch 3.2 GB、ffmpeg1.2 GB、ollama按模型规模。 - 历史回填 —— 同一签名跑过之后,用实测增量修正规则估算。
安装
两种装法:列表与插件市场用的是第一种,第二种用于改代码。
从 GitHub 安装(推荐)
在 DSH 里打开 插件 → + 添加插件,填:
github:wth373737/dsh-resource-guard
装完重启一次 DSH —— 替换已安装的包,只有重启后才会加载新的模块代。
从本地目录安装(开发用)
link: 安装会让 profile 直接指向那个目录,改完重启即生效,更新只需 git pull。所以别放在临时目录或会话工作区里。
- 克隆:
git clone https://github.com/wth373737/dsh-resource-guard.git - 在运行中的 profile 里,用绝对目录路径安装(插件管理器的
install_bundle,或让代理代劳)。 - 重启一次 DeepSeek Harness。
两种装法都可以用同一句确认:下一次请求的系统上下文里会出现资源预算区块(提交 / 显存 / 建议并发)。
配置
写在 bundle 的 cordis.patch.yml 的 config 下,由插件管理器重载。
| 键 | 默认 | 含义 |
|---|---|---|
enabled | true | 总开关 |
mode | ask | ask 询问 / deny 直接拒 / warn 只记录不拦 |
scope | heavy | heavy 只管网关工具 / all 全部工具 |
reserveGb | 6 | 判断物理内存压力时保留的余量 |
askRatio | 0.8 | 预计提交/上限 超过此比例就询问 |
denyRatio | 0.94 | 超过此比例直接拒绝 |
vramAskRatio | 0.85 | 显存询问线 |
vramDenyRatio | 0.97 | 显存拒绝线 |
minEstimateGb | 4 | 低于此估计不打扰 |
perWorkerGb | 3.2 | 计算安全并发时假定的单进程峰值;历史样本 ≥ 3 条后被实测中位数替代 |
breaker | true | 运行时监控 |
breakerRatio | 0.92 | 运行中提交占比警戒线 |
injectBudget | true | 是否注入预算区块 |
learn | true | 是否回填历史 |
autoTune | suggest | off / suggest(只生成依据文档)/ apply(同时把推荐值写进配置) |
tools | [] | 额外强制纳入门禁的工具名 |
debug | false | 同时写 stderr |
设置页
装上 bundle 之后,设置 → 资源门禁 会出现一个页面。
这个页面由 bundle 的客户端半边(dsh.client + client.js)注册,不是从 Host 配置自动生成的:设置里的每个页面都由客户端插件注册到 settings.section,只有 Host 半边的插件根本不会出现在那里——配置存在,但没有任何东西能显示或修改它。v0.3.0 补上客户端半边就是为了这件事。
页面显示实时预算(提交内存、显存、安全并发)、硬件推导建议(带「填入推荐值」),以及每个阈值的可编辑字段和说明。读写走一条带自定义头的同源路由(/resource-guard/api,要求 x-resource-guard-ui),Host 侧通过 ctx.configEditor 读取并按 Loader 的正常路径写入——和插件管理器用的是同一条通道,所以保存的值会落进 profile patch,和手改配置完全等价。settings 服务仅作为兜底。
装新版本后刷新一次页面——客户端 bundle 在启动时加载。
页面跟随系统语言(navigator.language):中文系统显示中文,其余英文。语言在页面加载时确定,改完系统语言需要刷新。
配置依据文档
插件在启动时实测本机硬件,写出 $DSH_HOME/resource-guard/recommendation.md:推荐阈值,以及每一条阈值来自哪项硬件属性。例如:
askRatio→ 0.78 —— 提交上限 66.9 GB ÷ 物理 62.9 GB = 1.06,分页文件只有 4 GB,几乎没有溢出空间,所以询问线压到 78%。
autoTune: suggest(默认)只生成这份文档;apply 会在硬件指纹变化后把上表写进配置,记录在 applied.json;off 完全跳过。
状态文件
$DSH_HOME/resource-guard/
history.json—— 每个签名的典型值(运行平均,用于估算)与峰值(用于告警)。可随时删,删了只是退回纯规则估算。recommendation.md—— 本机的硬件推导建议,每次启动重新生成。applied.json—— 最近一次autoTune: apply的硬件指纹。guard.log—— 最近运行日志,超过 512 KB 自动轮转。
已知边界
- 不透明命令仍然只能猜:并发数若由运行时计算(读配置、按 CPU 核数动态决定),静态分析看不到,靠运行时熔断兜底。
- 显存是慢采样:一次
Get-Counter要 1–1.5 秒,所以显存每 10 秒更新、内存每 3 秒,瞬时尖峰可能被漏掉。 - 没有硬杀:v1 只告警并请模型自行收敛。安全地只杀掉某次调用派生的进程需要进程树归属,误杀的代价比事故更大,留到 v2。
- Schema 加载是尽力而为:设置页依赖 Harness 自己的
schemastery;解析不到时插件照常运行,只是没有表单。 - 两个常驻 PowerShell 采样进程,各约 40–60 MB。
环境要求
带插件管理器的 DeepSeek Harness;Windows(采样依赖 GlobalMemoryStatusEx 与 WDDM 性能计数器)。
卸载
插件管理器里移除 dsh-plugin-resource-guard 这个 bundle,再删掉 $DSH_HOME/resource-guard/。
许可
MIT © 2026 wth373737,详见 LICENSE。
Похожие плагины
deepseek-harness
deepseek-ai/deepseek-harness
dsh-market
dsh-market/dsh-market
dsh-web (dsh-plugin-manager)
zhu1090093659/dsh-web
dsh-web
zhu1090093659/dsh-web