@Eason-22/dsh-autorestart avatar

@Eason-22/dsh-autorestart

DSH 插件:重启不打断工作——等所有对话空闲后优雅重启,重启后自动接着干。AI 生成、未人工审核,谨慎使用。

PluginAutomationAgent orchestration
Verification
L5 · Run tested
Verified
Risk
LowRisk
Health
Active
Trust
Gold

Install verification: runtime verified (L5) · tested on dsh 0.2.1-rc.2 · last checked Install verification details →

Install verification badge dsh 0.2.1-rc.2

What it doesAI

Restarts DSH only when every conversation is idle, then auto-resumes interrupted sessions and re-arms goals so work continues.

  • Waits until all sessions are idle before restarting, never mid-task
  • After restart, resumes interrupted sessions and re-arms persistent goals
  • Merges concurrent restart requests into one and wakes every requester
automationAIagent-orchestrationAI

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 dsh-autorestart
verified 0.2.1

Install method: npm · runtime verified (L5)

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

Declared DSH compatibilityfrom package.json · peerDependencies
package manifest not collected

package manifest not collected — the declared range is unknown.

Install compatibility matrix (rows = plugin versions〔latest 5〕, columns = dsh versions〔latest 5〕)
At-a-glance → evidence in the cards below
plugin \ dshdsh 0.2.1-rc.2
current
0.2.1✓
Install compatibility (by plugin version, dsh versions within)
Legend:✓ L5 runtime verified○ ecosystem plugin◼ ecosystem app✗ failed
0.2.1npm✓ 1 runtime-verified2026-10-22

dsh-autorestart

✓dsh 0.2.1-rc.2current2026-10-22L5 · runtime verified
Environment & artifact
  • • Node.js: v24.14.0
  • • DSH: 0.2.1-rc.2
  • • Plugin artifact: npm · dsh-autorestart@0.2.1
Levels (L1–L3 plugin-level static archive · L4/L5 newest real verdicts for this combination)
L1 static
Passed
L2 static
Passed
L3 static
Passed
L4 sandbox
Passed
L5 runtime
Passed
Install + runtime smoke verified in the dsh sandbox — installs and runs
L5 · runtime verified · valid for dsh 0.2.1-rc.2

ID: dsh-autorestart@0.2.1@dsh0.2.1-rc.2 · profile: l4-sandbox

Issued: 2026-10-22 · expires: 2026-10-29

View checks & evidence
L1 · Found✓ Passed

↳ carried from earlier record dsh-autorestart@9e57f2@dsh0.2.1-rc.2 (l3-fullscan)

Defaulted from L3✓ Passed
defaulted to passed — L3 install spec passed (full-registry L3 sweep policy)
L2 · Structured✓ Passed

↳ carried from earlier record dsh-autorestart@9e57f2@dsh0.2.1-rc.2 (l3-fullscan)

Defaulted from L3✓ Passed
defaulted to passed — L3 install spec passed (full-registry L3 sweep policy)
L3 · Install spec✓ Passed

↳ carried from earlier record dsh-autorestart@9e57f2@dsh0.2.1-rc.2 (l3-fullscan)

Install command parsed✓ Passed
Source: registry verification · Method: install spec parse · Captured 2026-10-22
DSH version declared✓ Passed
Source: registry verification · Method: DSH version declaration · Captured 2026-10-22
L4 · Install tested✓ Passed
Install executed in confined sandbox✓ Passed
Source: Windows ACL sandbox · Method: sandbox install run · Captured 2026-10-22
spec dsh-autorestart · channel npm package · outcome ran · exit 0
check took: 3.7s
Install succeeded✓ Passed
Source: Windows ACL sandbox · Method: sandbox install run · Captured 2026-10-22
spec dsh-autorestart · channel npm package · outcome ran · exit 0
check took: 3.7s
Installed artifact confirmed✓ Passed
Source: Windows ACL sandbox · Method: sandbox install run · Captured 2026-10-22
spec dsh-autorestart · channel npm package · outcome ran · exit 0
check took: 3.7s
Resolved version recorded✓ Passed
Source: Windows ACL sandbox · Method: sandbox install run · Captured 2026-10-22
spec dsh-autorestart · channel npm package · outcome ran · exit 0
check took: 3.7s
L5 · Run tested✓ Passed

active in loader tree (150 entries)

Sandbox install passed✓ Passed
Source: Windows ACL sandbox · Method: sandbox install run · Captured 2026-10-22
spec dsh-autorestart · install mode reused L4 scratch
check took: 642ms
Web boot ready✓ Passed
Source: DSH runtime · Method: web boot · Captured 2026-10-22
ready line dsh web: http://127.0.0.1:55161/?token=…
HTTP endpoint served✓ Passed
Source: DSH runtime · Method: HTTP request · Captured 2026-10-22
path / · HTTP 200 · took 642ms
Plugin active in inventory✓ Passed
Source: DSH runtime · Method: plugin inventory query · Captured 2026-10-22
verdict active · entries 150
match: include:dsh-autorestart · dsh-autorestart · phase active
phases: active 122 · null 28

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 official badges in your README to showcase security & install-test status:

dsh.so risk
`[![dsh.so risk](https://www.dsh.so/badge/dsh-autorestart.svg)](https://www.dsh.so/artifact/dsh-autorestart/)`
dsh 0.2.1-rc.2 install verified
`[![dsh.so install · dsh 0.2.1-rc.2](https://www.dsh.so/badge/install/dsh-autorestart@0.2.1-rc.2.svg)](https://www.dsh.so/artifact/dsh-autorestart/)`

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.

dsh-autorestartAutomated review · daily

Static low, dynamic medium — stricter wins → overall medium.

static low→dynamic medium→overall medium
0 critical·0 high
mediumoverall · medium
1static scan · vet primaryvet AST static analysis (npm artifact or git source)
plugin versionv0.2.1dsh manifest
scanned version0.2.1npm2026-10-22
current version0.2.1same version
vet verdictclean · npm

Vet is an AST static scan (npm artifact or git source).

show static findings (6)
  • criticalspawnShell · shelllib/restart.js:351

    const child = spawn(text, { shell: true, detached: true, windowsHide: true, stdio: 'ignore' });

  • warningnetIp · networkindex.js:566

    return 'http://127.0.0.1:' + port + resolveHealthPath(config) + '?gate=1'

  • warningfsWrite · fileslib/checkpoint.js:431

    fs.writeFileSync(fd, payload, 'utf8');

  • warningshellExec · shelllib/restart.js:295

    const stdout = execFileSync(wmiCommand.file, wmiCommand.args, {

  • warningshellExec · shelllib/restart.js:322

    execFileSync('schtasks', schtasksCreate, execOptions);

  • warningshellExec · shelllib/restart.js:323

    execFileSync('schtasks', schtasksRun, execOptions);

2model reads redacted source context · never executes plugin code
highest grademedium
merged intooverall medium
audited at2026-10-22
findings0 critical·0 high1 medium / 2 low hidden
show 3 audit findings
  • mediumOperator-configured restartCommand is spawned through a shell without per-invocation confirmationlib/restart.js:351

    When restartMode is 'command' and restartCommand is non-empty, executeRestart() expands the template (buildCustomCommand) and runs it detached through a real shell as the Host user, triggered automatically by a model tool call (continuum_restart) with no confirmation step. The command text is operator-supplied, not tool-arg controlled, and the default (restartMode 'app', empty restartCommand) never reaches this path, so this is a capability/config risk rather than an injection from untrusted input.

  • lowRestart helper kills every process named "DeepSeek Harness.exe" machine-wide, not just this hostlib/restart.js:116

    taskkill /IM with the default image name plus /T /F terminates every process with that image name on the machine, including a second/dev instance or a window a user opened deliberately; combined with the ungated fallback after more than 20 gate aborts, the kill can also land while another conversation is mid-turn, losing in-flight work.

  • lowGate helper .cmd is written next to the checkpoint and later executed via schtasks/WMIlib/restart.js:283

    A .cmd file (restart-<generation>.cmd) is materialized under path.dirname(stateFile) and then run as `cmd.exe /c "<script>"` through schtasks/WMI. The content is generated from appExePath and the health URL, so it is not attacker-authored, but anyone able to write into the (operator-configurable) checkpoint directory could substitute the file between write and execution and have their content run with the Host's privileges.

Restart path is genuinely gated (idle gate, loopback re-check, rollback caps); residual risk is an operator-configured command spawned through a shell and a broad taskkill/write-execute helper.

Rule: the dynamic audit can only raise risk, never lower it (raise-only, never an exemption); a wrong raise is withdrawn by human review and the tier falls back to the static verdict. Overall verdict = the stricter of the two layers.

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 2026-10-22 · activity: Active

• Repo created: 2026-10-22

• Stars: ★ 0 · Forks: 0

• Health: Active — committed within last 30 days

Source

GitHub: github.com/Eason-22/dsh-autorestart

Was this page helpful?