dsh-grantwrite-patch
Windows sandbox ACL-repair plugin for DeepSeek Harness (dsh) — fix the grantWrite failure (SetNamedSecurityInfoW Win32 5) and retry the command invisibly
456 results
Windows sandbox ACL-repair plugin for DeepSeek Harness (dsh) — fix the grantWrite failure (SetNamedSecurityInfoW Win32 5) and retry the command invisibly
Per-preset tool visibility for DeepSeek Harness: apply ctx.tools.restrict() on an agent's own scope at agent/created, so a preset exposes only the tools it should while the heavy providers stay mounted.
DSH plugin that bundles the upstream reverse-skill corpus (88 skills) as a skill provider and tracks upstream commits.
dsh plugin: let workspace-write sandboxed commands write the Flutter SDK, pub cache, Dart/Flutter settings, Gradle, and Android directories.
Audible alert for the dsh web GUI: plays a distinct sound the moment the harness raises a prompt that is waiting for you — approval panels AND ask-user-question panels — with synthesized per-seam voices or your own uploaded audio, pushed through a limiter so it can exceed 100%
DeepSeek Harness plugin: drop sandbox_permissions/justification fields that cannot widen the current sandbox mode, so redundant escalations (common with GPT models) no longer fail with "not strictly wider".
End-to-end encrypted remote access to DeepSeek Harness and experimental Codex workspaces from desktop, web, and Android, with dsh-TUI Host support.
DSH 审批弹窗 AI 注释器:在模型请求权限外操作时,用模型自己的推理生成「这段代码做什么 / 想达成什么 / 风险评估 / 两个选项各自意味着什么」,让用户带着依据做决定,而不是瞪着一行代码点按钮。
查密钥与证书多久没换了、有没有轮换制度 —— 长期不换的密钥往往没人注意,却是实打实的风险
检查告警链路:有没有告警、发到哪、有没有把敏感信息一起发出去
查查阅/复制/删除权有没有真的实现 —— 个保法第 45/47 条是明文要求
查应用里有没有展示 ICP/APP 备案号、隐私政策入口、运营者信息 —— 都是法定义务
Route QQ private messages into the AI team sessions you already run in DeepSeek Harness -- same session, same memory, same permissions. No second QQ gateway, no copy-pasted system prompt.
开源许可合规检查:扫描依赖树,找出 GPL/AGPL 传染、许可未知等风险 —— 发布前、开源前、被审计时跑一遍。
System-prompt surface drift gate for DeepSeek Harness: baselines every section contributed to the assembled system prompt and warns loudly when a plugin or preset upgrade silently adds, removes, or rewrites what the model is told
Agent governance suite for DeepSeek Harness (enterprise companion): policy-based tool gating (allow/deny/ask), structured JSONL audit trail, and per-agent token quotas against the host token meter — state under $DSH_HOME/gov
JumpServer asset lookup for DeepSeek Harness: query assets via conversation using AccessKey/SecretKey signed requests
DSH permission-gate for the DeepSeek Harness 0.1.5 line: a single self-sufficient, deterministic-first, fail-closed gate covering P0 hard-deny -> P1 session grant -> P2 static rule (allow/deny) chain -> P3 optional LLM semantic classifier -> P4 ask, with
Zero-dependency static + sandbox smoke detector for DeepSeek Harness (dsh) plugins: package-structure gates (R), cordis contract scans (K), keyless-headless sandbox smoke (D), and ecosystem-listing checks (CC).
Patch-style file editing for DeepSeek Harness: one apply_patch tool accepting git/unified diff (default) and Codex apply_patch syntax (opt-in), with all-or-nothing application, read-before-write intent dance, sandbox-aware delete/move, (declarative volat
Permission matrix for DeepSeek Harness: 3 sandbox modes x 4 approval strategies = 9 selectable presets, with a global default, a separate LLM-robot default, four-tier risk policies, password-gated approval for high-risk operations, audit log and git checkpoint.
Local non-generative decision layer for DeepSeek Harness: the phocinae_ask / phocinae_gate tools backed by a Phocinae-Largha-150M-v1 server, plus a fail-closed approval gate on tools/pre-execute with a calibrated-confidence escalation signal
Second-model AI auto-review for DeepSeek Harness approval requests: a read-only reviewer subagent decides allow/deny on the approval answerer chain, with fail-closed fallback and full session-log audit.
Declarative Claude Code-style permission rules plus a Codex-style process-level network policy for DeepSeek Harness: ordered allow/deny/ask rules with tool-name, argument (glob/regex), workspace-path, and network-target (domain/ip/port/scheme) matching on