⚠️ This repository has been renamed: BOWLUNA/dsh-multi-instance — click to visit the new location. Everything below is preserved as originally indexed.
DSH 多开桌面客户端:把 dsh web 装进原生 webview 窗口,按 URL 接入多个实例 —— 本机、WSL2、或服务器上的远程 DSH 都能接。多窗口自由排布与缩放,多实例之间互不干扰,每格独立浏览器会话。|Desktop client and webview shell for DeepSeek Harness: multi-instance, multi-window panes that attach to any number of DSH instances by URL - local, WSL2 or remote - one isolated browser session per pane.
Install verification: Ecosystem plugin: installs via dsh plugin add but declares no dsh.bundle and mounts no plugin capability — the L1-L5 install ladder does not rate it; the security scan still runs.Install verification details →
Data as of (UTC)
What it doesAI
Save, one-click switch, import/export and share multiple system prompts and personas for DeepSeek Harness.
- Store multiple system prompts and personas in one hub
- Switch between saved prompts with one click
- Import, export and share prompt collections
Capability tags are AI-inferred or rule-extracted from the README (see badge) — vocabulary-limited, not install-verified. AI = inferred; README = literal keyword from the readme.
Trends
Four lines, four questions: total stars, stars gained this week, cumulative downloads, downloads this week. The GitHub pair comes from real-read sampling (a rotating daily shard); the npm pair from every snapshot.
3 samples (roughly every 5 days), 3 of them reconstructed (hollow dots); sampled points, not a daily series
09-21 → 10-07 · 3 points
Each bar = the delta between two samples 5–14 days apart; negative bars point down. Faint bars include reconstructed points
Only 2 points (< 3): no chart yet
10-02 → 10-07 · 2 points
Reconstructed backwards from the current total using daily downloads (gaps are not interpolated)
Only 0 points (< 3): no chart yet
no data yet
The snapshot’s last-week rolling window (official npm + npmmirror combined)
Only 0 points (< 3): no chart yet
no data yet
Installation
dsh plugin --profile web add github:BOWLUNA/dsh-prompt-hubInstall method: GitHub · runtime test inconclusive (L5 unknown) · verified on dsh 0.1.7-rc.2
Verification & Compatibility○ Ecosystem plugin
Grouped by plugin version: each card lists the dsh versions actually tested against that artifact version and their install verdicts (L5 wins). Untested combinations are omitted — absence means untested, never assumed compatible. L1–L3 are a plugin-level static archive (independent of the dsh version — the same verdict on every dsh row): they sit in each row's level grid next to L4/L5, marked "static", while their evidence is stored once per plugin inside that row's "checks & evidence" instead of being duplicated per version.
External credentials (plugin-level): none detected
○ The author declares no dsh version requirement — install is not blocked by version.
| plugin \ dsh | dsh 0.1.7-rc.2 | dsh 0.1.7-rc.1 | dsh 0.1.6-alpha.2 |
|---|---|---|---|
| 9d2c2d | · | · | · |
| 844b02 | ○ | ○ | ○ |
9d2c2dlocal2026-09-18
Ecosystem plugin: installs via dsh plugin add but declares no dsh.bundle and mounts no plugin capability — its level reads L5 · Ecosystem plugin.
844b02github○ 3 ecosystem plugin2026-09-25
https://github.com/BOWLUNA/dsh-prompt-hub
○dsh 0.1.7-rc.22026-09-25Ecosystem plugin
- • Node.js: v24.14.0
- • DSH: 0.1.7-rc.2
- • Plugin artifact: github · https://github.com/BOWLUNA/dsh-prompt-hub (version unstated)
ID: dsh-prompt-hub@844b02@dsh0.1.7-rc.2 · profile: l4-sandbox
Issued: 2026-09-25 · expires: 2026-10-02
View checks & evidence
not in scope for web smoke; see prior standard-profile records
not in scope for web smoke; see prior standard-profile records
↳ carried from earlier record dsh-prompt-hub@9d2c2d@dsh0.1.6-alpha.1 (l3-fullscan)
not present in web-profile loader inventory
L1–L3 are plugin-level static conclusions (baseline, identical across dsh versions): evidence is stored once per plugin.
○dsh 0.1.7-rc.12026-09-23Ecosystem plugin
- • Node.js: v24.14.0
- • DSH: 0.1.7-rc.1
- • Plugin artifact: github · https://github.com/BOWLUNA/dsh-prompt-hub (version unstated)
ID: dsh-prompt-hub@844b02@dsh0.1.7-rc.1 · profile: l4-sandbox
Issued: 2026-09-23 · expires: 2026-09-30
View checks & evidence
not in scope for web smoke; see prior standard-profile records
not in scope for web smoke; see prior standard-profile records
↳ carried from earlier record dsh-prompt-hub@9d2c2d@dsh0.1.6-alpha.1 (l3-fullscan)
This level was not executed in this scan.
L1–L3 are plugin-level static conclusions (baseline, identical across dsh versions): evidence is stored once per plugin.
○dsh 0.1.6-alpha.22026-09-18Ecosystem plugin
- • Node.js: v24.14.0
- • DSH: 0.1.6-alpha.2
- • Plugin artifact: github · https://github.com/BOWLUNA/dsh-prompt-hub (version unstated)
ID: dsh-prompt-hub@844b02@dsh0.1.6-alpha.2 · profile: l4-sandbox
Issued: 2026-09-18 · expires: 2026-09-25
View checks & evidence
not in scope for web smoke; see prior standard-profile records
not in scope for web smoke; see prior standard-profile records
↳ carried from earlier record dsh-prompt-hub@9d2c2d@dsh0.1.6-alpha.1 (l3-fullscan)
not present in web-profile loader inventory
L1–L3 are plugin-level static conclusions (baseline, identical across dsh versions): evidence is stored once per plugin.
Author badges
Embed the dsh.so badges in your README to show this plugin's automated security-scan and install-test status:
`[](https://www.dsh.so/artifact/dsh-prompt-hub/?ref=badge)`dsh.so is an independent community project, not affiliated with or endorsed by DeepSeek AI. Badges are maintained by dsh.so; embed links carry a ?ref= attribution parameter.
Security Report
Two automated layers, neither a manual review: a static scan (per-plugin rule checks) and a model-driven dynamic security audit that covers only the subset static-flagged at warn or above and never executes plugin code.
Static low; dynamic medium (authoritative when present) → overall medium.
—d8eb824eaf31git2026-09-26v0.5.0version moved onclean · gitVet is an AST static scan (npm artifact or git source).
show static findings (12)
- warningchildProcess · code-exec
bin/cli.js:76const child = spawn(electronPath, args, { stdio: 'inherit', env, windowsHide: false }); - warningchildProcess · code-exec
src/main/discovery.js:46child = spawn(command, args, { - warningchildProcess · code-exec
src/main/discovery.js:145child = spawn('netstat', ['-ano'], { windowsHide: true }); - warningchildProcess · code-exec
src/main/discovery.js:198child = spawn('wsl.exe', ['-l', '-q'], { windowsHide: true }); - warningchildProcess · code-exec
src/main/installer.js:46child = spawn(command, args, { - warningchildProcess · code-exec
src/main/installer.js:222child = spawn(command, args, { - warningchildProcess · code-exec
src/main/instances.js:88child = spawn('wsl.exe', args, { windowsHide: true }); - warningchildProcess · code-exec
src/main/instances.js:316child = spawn(built.command, built.args, { - warningchildProcess · code-exec
src/main/instances.js:343const child = spawn(built.command, built.args, { - warningfsWrite · files
src/main/main.js:69logStream = fs.createWriteStream(path.join(dir, 'main.log'), { flags: 'a' }); - warningfsWrite · files
src/main/main.js:416fs.writeFileSync(file, img.toPNG()); - warningfsWrite · files
src/main/store.js:48fs.writeFileSync(tmp, JSON.stringify(this.data, null, 2), 'utf8');
show 4 audit findings
- mediumTLS certificate verification disabled when probing instance URLs
src/main/instances.js:211probe() is the reachability check for an instance URL, and resolveUrl()/applyToken() append the dsh access token to that URL (`url + '?token=' + encodeURIComponent(tk)`), so the token is placed into the request path. With certificate validation off, any network attacker (or a wrong/captive host) can answer the probe with an arbitrary certificate, learn that a token-bearing URL is being polled, and read the request; the probe's own `looksLikeDsh` verdict is also meaningless because the server identity is never verified.
- mediumCaller-supplied dir/registry/version are pasted unescaped into cmd.exe and bash command lines
src/main/installer.js:197install({dir, registry, distro, user, version}) builds a whole shell line by interpolation and then runs it via `wsl.exe ... bash -lc <script>` or `cmd.exe /d /s /c <cmd>` with windowsVerbatimArguments. A value containing a quote plus `&`, `|`, `;` or `$(...)` escapes the quoted argument and executes arbitrary commands with the user's privileges; `version` flows through specFor() into the package spec, so a value containing spaces can append arbitrary npm flags (e.g. another --registry). The same interpolation appears in verify() (line 246) and binPath-derived paths. Precondition: the values c
- lowPersisted launcher command and log-file path are executed as shell commands
src/main/instances.js:269Whatever sits in config.json's `launcher.command` (and the `logFile` path, which is interpolated into a bash command) is executed via `bash -lc` or `cmd.exe /c` as soon as that instance is started — including automatically when `autostart` is set. appdata.js migrates a config.json out of legacy folders (%APPDATA%\dsh-shell, 'DSH套壳', …) whenever that folder has more panes, so a config planted in a no-longer-used directory can be pulled in and executed. This is partly the intended feature, and it requires local write access to the user's AppData config (i.e. the attacker already runs as that use
- lowPreload bridge exposes arbitrary JS execution inside any webview by id
src/preload/preload.js:40The comment says it is open only for automation runs (--eval/--shot), but src/main/main.js was not harvested, so that gate could not be verified. If the handler does not re-check the flag in the main process, any code running in the host renderer can execute arbitrary JavaScript in any pane webview — including panes attached to a remote DSH instance whose session token/session is live. No obfuscation or extra argv/command-line processing was found elsewhere, so this is a hardening gap, not a demonstrated exploit.
Electron shell for local dsh instances: no backdoor found, but TLS verification is disabled in probes and caller-supplied paths/registries go unescaped into cmd/bash command lines.
Rule (2026-10-28): when a dynamic conclusion exists it is authoritative (it may raise or lower the level); the static scan tier (vet primary, dsh backstop) applies only when there is none (not queued / audit failed / withdrawn). The audit never writes the exemption list (exemptions are manual only); a wrong conclusion is withdrawn by human review and the tier falls back to the static scan tier.
Disclaimer: the static scan and the dynamic security audit are both automated analysis, not a security guarantee. Always review what you install.
Activity
Last commit 2026-10-26 · activity: Active
• Repo created: 2026-09-17
• Stars: ★ 0 · Forks: 0
• Health: Active — committed within last 30 days
Source
Author README
View on GitHub ↗The author's original README from the repository — fetched live, unedited (the English page loads the default README).
Frequently asked questions
What is dsh-prompt-hub?
dsh-prompt-hub is an ecosystem plugin listed on dsh.so: it installs with dsh plugin add but declares no dsh.bundle and mounts no plugin capability, so installing it adds no plugin functionality — use its source repository directly.
Is dsh.so or this plugin page official?
dsh.so is an independent community registry and is not affiliated with, endorsed by, or sponsored by DeepSeek AI.
How do I install dsh-prompt-hub?
Run dsh plugin --profile web add github:BOWLUNA/dsh-prompt-hub in a DeepSeek Harness (dsh) environment. Note this is an ecosystem plugin: it declares no dsh.bundle and mounts no plugin capability, so installing it adds no plugin functionality — use the source repository directly. The security scan still runs.
How does the security scan relate to a manual audit?
Both layers are automated, not a manual audit. The static scan is heuristic analysis of the repository source, so it can produce false positives and miss context. The model-driven dynamic security audit only covers plugins static-flagged at warn or above, never executes plugin code, and its conclusion is authoritative when there is one (it may raise or lower the static tier); exemptions stay manual. Neither layer is an endorsement: review the repository and the permissions it asks for before installing, together with the verification level and health signals.
How fresh is the data on this page?
The data on this page is as of 2026-10-26 (UTC) — the most recent valid timestamp among the registry refresh, install verification, security scan and dynamic audit. Build time is never used as a data date.
