Install verification: sandbox install failed (L4) · tested on dsh 0.2.1-rc.5 · last checked Install verification details →
Data as of (UTC)
What it doesAI
Crash-surviving background job manager for DeepSeek Harness with filesystem as source of truth.
- Cross-restart job adoption
- Autopsy reports for crashes
- Sandboxed execution and event sourcing
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.
Installation
dsh plugin --profile web add github:Wang-Lin-Chang/dsh-witness#v0.2.1Install method: GitHub · sandbox install failed (L4) · verified on dsh 0.2.1-rc.5
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, so one plugin is re-read roughly every 5–14 days, sparser than npm; the npm pair has a value on every snapshot day. The two cumulative panels are smoothed curves with no sample markers; the two weekly panels keep their raw bars and add a 4-week moving-average trend line that strips single-week noise down to the direction of travel.
7 samples (roughly every 7 days), 3 of them reconstructed (dashed runs); a smoothed curve through every sample — not a daily series
09-22 → 11-14 · 7 points
Bars = the delta between two samples 5–14 days apart (negative points down, faint bars contain reconstructed points); line = 4-week moving-average trend
10-06 → 11-14 · 3 points
Reconstructed backwards from the current total using daily downloads (gaps are not interpolated)
Only 0 points (< 3): no chart yet
no data yet
Bars = the snapshot’s last-week rolling window (official npm + npmmirror combined); line = 4-week moving-average trend — the shape of weekly demand, not a user count
Only 0 points (< 3): no chart yet
no data yet
Verification & Compatibility
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
Based on 1 requirement(s) across 1 @deepseek-ai/dsh-* packages. raw ranges
0.1.0-rc.6 ×1
✗ Running dsh 0.2.2-rc.5 is outside the declared range — the official install gate rejects it (incompatible-version).
To force-install, grant an exact-version exemption: dsh plugin allow-version dsh-witness@0.2.1 0.2.2-rc.5 (the exemption lives in the local profile; this site does not assume one)
Deciding channel: no engines.dsh declared, so every dsh* peer is checked (workspace:^ | ~ | * counts as the running version).
Whether it actually runs is a separate question → see the measured matrix below (L1–L5).
| plugin \ dsh | dsh 0.2.1-rc.5 | dsh 0.2.1-rc.3 | dsh 0.2.1-rc.2 | dsh 0.2.1-alpha.4 | dsh 0.2.1-alpha.2 |
|---|---|---|---|---|---|
| v0.2.1 | ✗ | ✗ | ○ | ○ | ○ |
v0.2.1github2 failed2026-11-14
https://github.com/Wang-Lin-Chang/dsh-witness
✗dsh 0.2.1-rc.52026-10-31L4 · failed
- • Node.js: v24.14.0
- • DSH: 0.2.1-rc.5
- • Plugin artifact: github · https://github.com/Wang-Lin-Chang/dsh-witness#v0.2.1
ID: dsh-witness@v0.2.1@dsh0.2.1-rc.5 · profile: l4-sandbox
Issued: 2026-10-31 · expires: 2026-11-07
View checks & evidence
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 0.2.1-rc.32026-10-26L4 · failed
- • Node.js: v24.14.0
- • DSH: 0.2.1-rc.3
- • Plugin artifact: github · https://github.com/Wang-Lin-Chang/dsh-witness#v0.2.1
ID: dsh-witness@v0.2.1@dsh0.2.1-rc.3 · profile: l4-sandbox
Issued: 2026-10-26 · expires: 2026-11-02
View checks & evidence
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 0.2.1-rc.22026-10-20not tested
- • Node.js: v24.14.0
- • DSH: 0.2.1-rc.2
- • Plugin artifact: github · https://github.com/Wang-Lin-Chang/dsh-witness#v0.2.1
ID: dsh-witness@v0.2.1@dsh0.2.1-rc.2 · profile: l4-sandbox
Issued: 2026-10-20 · expires: 2026-10-27
View checks & evidence
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 0.2.1-alpha.42026-10-18not tested
- • Node.js: v24.14.0
- • DSH: 0.2.1-alpha.4
- • Plugin artifact: github · https://github.com/Wang-Lin-Chang/dsh-witness#v0.2.1
ID: dsh-witness@v0.2.1@dsh0.2.1-alpha.4 · profile: l4-sandbox
Issued: 2026-10-18 · expires: 2026-10-25
View checks & evidence
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 0.2.1-alpha.22026-10-11not tested
- • Node.js: v24.14.0
- • DSH: 0.2.1-alpha.2
- • Plugin artifact: github · https://github.com/Wang-Lin-Chang/dsh-witness#v0.2.1
ID: dsh-witness@v0.2.1@dsh0.2.1-alpha.2 · profile: l4-sandbox
Issued: 2026-10-11 · expires: 2026-10-18
View checks & evidence
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 0.2.0-rc.22026-10-09not tested
- • Node.js: v24.14.0
- • DSH: 0.2.0-rc.2
- • Plugin artifact: github · https://github.com/Wang-Lin-Chang/dsh-witness#v0.2.1
ID: dsh-witness@v0.2.1@dsh0.2.0-rc.2 · profile: l4-sandbox
Issued: 2026-10-09 · expires: 2026-10-16
View checks & evidence
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 0.2.0-rc.12026-09-30not tested
- • Node.js: v24.14.0
- • DSH: 0.2.0-rc.1
- • Plugin artifact: github · https://github.com/Wang-Lin-Chang/dsh-witness#v0.2.1
ID: dsh-witness@v0.2.1@dsh0.2.0-rc.1 · profile: l4-sandbox
Issued: 2026-09-30 · expires: 2026-10-07
View checks & evidence
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.22026-09-27not tested
- • Node.js: v24.14.0
- • DSH: 0.1.7-rc.2
- • Plugin artifact: github · https://github.com/Wang-Lin-Chang/dsh-witness#v0.2.1
ID: dsh-witness@v0.2.1@dsh0.1.7-rc.2 · profile: l4-sandbox
Issued: 2026-09-27 · expires: 2026-10-04
View checks & evidence
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 0.1.6-alpha.22026-09-20not tested
- • Node.js: v24.14.0
- • DSH: 0.1.6-alpha.2
- • Plugin artifact: github · https://github.com/Wang-Lin-Chang/dsh-witness#v0.2.1
ID: dsh-witness@v0.2.1@dsh0.1.6-alpha.2 · profile: l4-sandbox
Issued: 2026-09-20 · expires: 2026-09-27
View checks & evidence
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 0.1.6-alpha.12026-09-17not tested
- • Node.js: v24.14.0
- • DSH: 0.1.6-alpha.1
- • Plugin artifact: github · https://github.com/Wang-Lin-Chang/dsh-witness#v0.2.1
ID: dsh-witness@v0.2.1@dsh0.1.6-alpha.1 · profile: l4-sandbox
Issued: 2026-09-17 · expires: 2026-09-24
View checks & evidence
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 0.1.5-rc.22026-09-11not tested
- • Node.js: v24.14.0
- • DSH: 0.1.5-rc.2
- • Plugin artifact: github · https://github.com/Wang-Lin-Chang/dsh-witness#v0.2.1
ID: dsh-witness@v0.2.1@dsh0.1.5-rc.2 · profile: l4-sandbox
Issued: 2026-09-11 · expires: 2026-09-18
View checks & evidence
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 0.1.3-alpha.22026-09-08not tested
- • Node.js: v24.14.0
- • DSH: 0.1.3-alpha.2
- • Plugin artifact: github · github:Wang-Lin-Chang/dsh-witness#v0.2.1#v0.2.1
ID: dsh-witness@v0.2.1@dsh0.1.3-alpha.2 · profile: l5-web-smoke
Issued: 2026-09-08 · expires: 2026-12-07
View checks & evidence
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 0.1.3-alpha.12026-09-07not tested
- • Node.js: v24.14.0
- • DSH: 0.1.3-alpha.1
- • Plugin artifact: github · https://github.com/Wang-Lin-Chang/dsh-witness#v0.2.1
ID: dsh-witness@v0.2.1@dsh0.1.3-alpha.1 · profile: l4-sandbox
Issued: 2026-09-07 · expires: 2026-09-14
View checks & evidence
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.
○prehistory2026-10-04not tested
- • Node.js: v24.14.0
- • DSH: unknown
- • Plugin artifact: github · https://github.com/Wang-Lin-Chang/dsh-witness#v0.2.1
ID: dsh-witness@v0.2.1@dshunknown · profile: l4-sandbox
Issued: 2026-10-04 · expires: 2026-10-11
View checks & evidence
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.
Only the latest 3 published versions are shown; source-snapshot (commit hash) groups, unstated groups and older versions are omitted — the full archive stays in the data layer (snapshot/unstated groups appear only when a plugin has no published version).
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-witness/?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.
v0.2.1v0.2.1git2026-11-19v0.2.1same versionclean · npmVet is an AST static scan (npm artifact or git source).
show static findings (12)
- warningchildProcess · code-exec
detach-runner.cjs:34const r = require('node:child_process').spawnSync('powershell', ['-NoProfile', '-Command', `[int](Get-Date -Date (Get-Process -Id ${process. - warningchildProcess · code-exec
detach-runner.cjs:110const guard = spawn('powershell', ['-NoProfile', '-NonInteractive', '-Command', guardPs], { - warningchildProcess · code-exec
detach-runner.cjs:126try { require('node:child_process').spawnSync('powershell', ['-NoProfile', '-NonInteractive', '-Command', ps], { windowsHide: true, timeout: - warningfsWrite · files
detach-runner.cjs:40fs.writeFileSync(lockFile, `${process.pid}:0`, { flag: 'wx' }) // wx = O_EXCL 独占;lock 立即存在(收养观测窗口) - warningfsWrite · files
detach-runner.cjs:44if (lockCreated) { try { myStartSec = procStartSec(); fs.writeFileSync(lockFile, `${process.pid}:${myStartSec}`) } catch {} } // startSec - warningfsWrite · files
detach-runner.cjs:50try { fs.writeFileSync(exitFile, '', { flag: 'wx' }) } catch {} // 预创建 exit.txt(防任务抢建伪造;wx 防覆盖已有) - warningchildProcess · code-exec
lib/WitnessJobRegistry.js:319const r = spawnSync('powershell', ['-NoProfile', '-Command', `[int](Get-Date -Date (Get-Process -Id ${pid}).StartTime.ToUniversalTime() -UFo - warningchildProcess · code-exec
lib/WitnessJobRegistry.js:396spawn(process.execPath, [runnerPath, d, outFile, exitFile, cmdB64], { detached: true, stdio: 'ignore', windowsHide: true }); - warningfsWrite · files
lib/WitnessJobRegistry.js:185fs.writeFileSync(path.join(d, 'autopsy.json'), JSON.stringify({ - warningfsWrite · files
lib/WitnessJobRegistry.js:196fs.writeFileSync(path.join(d, 'state', 'done'), String(code)); - warningfsWrite · files
lib/WitnessJobRegistry.js:215fs.writeFileSync(tmp, ''); - warningchildProcess · code-exec
lib/spawnDirCleanup.js:16const r = spawnSync('powershell', ['-NoProfile', '-NonInteractive', '-Command', ps], { timeout: 8000, windowsHide: true });
show 3 audit findings
- mediumJob commands execute with full user privileges; the ACL 'sandbox' only protects job evidence files
detach-runner.cjs:181spec.command is base64-decoded (line 28) and handed to PowerShell as the harness OS user with no restricted token, AppContainer or job object. denyEvidence only adds deny ACEs to outFile, lockFile, exitFile and CreateFiles on spawnDir, so the child can read, write and delete anything else the user can reach (including arbitrary network/registry operations). Any consumer that treats this as an untrusted-code boundary (the package advertises "sandboxed execution") has a false sense of isolation; the in-code comment 'honest boundary ... arbitrary code level requires a restricted token (v6)' confi
- lowkill() terminates the PID read from the job lock without verifying process start time
lib/WitnessJobRegistry.js:452lockAlive() validates pid:startSec and rejects PID reuse, but kill() only regex-extracts the pid and calls process.kill on it. A job process (which runs as the same user and can, per the plugin's own documented boundary, self-rescue the lock ACL via P/Invoke) can write another PID into lock so that a subsequent kill(id) terminates an unrelated process; PID reuse without malice could also hit the wrong process. Impact is limited because such a task already holds arbitrary code execution as the same user.
- lowCaller-supplied spec.kind is used unsanitised in the job directory path
lib/WitnessJobRegistry.js:262The job id is built as `${kind}-${count}` (line 258-259) with no validation, and taskDir() joins it onto jobsRoot. A kind containing path separators (e.g. '..\\..\\target') normalises out of jobsRoot, so start() will mkdir/write spec.json there and the runner will create lock/out.log/exit.txt and apply/restore ACLs in that directory. Impact is low because the same caller already supplies the executed command; the escape adds no capability beyond it.
No injection found; job commands run with full user rights because the ACL 'sandbox' only protects job evidence files, and kill() trusts the lock PID without a start-time check.
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 2026-08-26 · activity: Stable
• Repo created: 2026-08-16
• Stars: ★ 0 · Forks: 0
• Health: Stable — committed within 90 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-witness?
dsh-witness is a DeepSeek Harness (dsh) plugin listed on dsh.so — an independent community registry that indexes public source repositories and publishes automated security-scan and install-verification results for each entry.
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-witness?
Run dsh plugin --profile web add github:Wang-Lin-Chang/dsh-witness#v0.2.1 in a DeepSeek Harness (dsh) environment. The command is parsed from the plugin manifest and its install spec is verified (L3); runtime sandbox install testing may not have run yet.
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 2026-11-19 (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.
