by taoshi1999

DeepSeek Harness plugin for agent workspace hygiene: artifact value assessment, metadata indexing, and auditable cleanup.

by onclaw-dev

Tencent IMA is an excellent knowledge base application, but the skill version they provide only offers keyword-based retrieval based on file titles for public knowledge base queries, which left me speechless for a while. To make up for this knowledge base retrieval capability in harness, I iterated on the corresponding dsh version based on tencent-ima-copilot-mcp.

by huanlin-labs

DSH plugin: default-deny directory authorization desk (isolation gate) with roster, groups, audit log and agent self-check

by GooDAnDReaDY

Sanitizes leaked DeepSeek DSML closing tags from model text streams.

by MoonlitDropOfBlood

An independent approval subagent judges every sandbox escalation, with a configurable model and an audit log.

by john-walks-slow

Per-agent model switching for DeepSeek Harness: set_model / list_models agent tools with provider whitelist and token-budget guardrails, plus plan-mode model staging and cache-safe active-model contex

by cdxiaodong

Agent security guardrail: intercepts and audits every tool call, requiring human confirmation on sensitive operations.

by 2424830124

Connect the workbuddy backend to dsh, allowing you to use workbuddy's models in dsh (consuming workbuddy credits)

by watersxya

AI compiled pure-text novel workbench for DSH: outline import, Daozang/Great World setting, volume and chapter planning, chapter-by-chapter AI generation of 3000-4000 characters, nine-dimensional review, whole-book quality inspection, full-book export. AI novel writing workbench (text-only).

by ateen18

Pre-install static security review and runtime guard for dsh plugins: deobfuscation decoding, supply-chain checks, web one-click review/install/uninstall, and optional runtime tool-call guard.

by azazo1

Protect declared workspace subpaths such as .git from writes, and optionally grant extra writable roots under workspace-write.

by sueqet

Cross-session TODO board as a draggable floating panel: tasks are grouped by working directory and reorderable by drag, each one carries a run mode (remind only, continue in the current session, or op

by StyxNether

Auto-Approval: a middle permission tier for DeepSeek Harness between workspace-write and danger-full-access, auto-approving harmless commands and trusted-area targets

by omdsh-dev

microsandbox support

by Plocr

DeepSeek Harness (dsh) plugin: publishes a Command Code subscription (GOAT / Pro / Max) as llm-pi-ai provider routes, with account usage and web search on the same credential.

by wulun811

Trust pipeline for deepseek-harness plugins: deterministic static scan (11 rules) + LLM-driven audit protocol + two-part scorecard, with optional runtime guard (T1 sentinel / T2 fs & child_process hoo

by bescriptkiddie

WeChat Official Account source collector: QR-login credential profiles, rate-limit-safe crawling with cooldown, RSS second water source, and secret-safe handoff to ContentStudio.

by bufan528

Obsidian-compatible long-term memory vault plugin for DeepSeek Harness (DSH): local Markdown storage, sidebar management page, token-budgeted recall, durable journal outbox.

by SiriLee

Per-edit approval gate for DeepSeek Harness — red/green line diff review before write/edit/stream operations · DSH plugin: per-operation approval gate, red/green line-level diff review before write/edit/stream execution

by drscrewdriver

P0–P4 deterministic-first permission gate with permissive tier, learning sedimentation and approval-history UI; hard-deny credentials and protected paths, auto-allow internal read-only tools.

by 940842546

Claude Code-style permission rules engine: hard/deny/ask/allow tiers with a hard tier above full access, workspace-scoped rules, wildcard path protection, and a visual staged editor; rules persist in

by cyanseek

Deterministic fault injection and autonomous resilience tests for DeepSeek Harness tools

by chenbin-dev

Import local Claude, Codex, Grok, Gemini, Copilot, OpenCode, and CC Switch configurations. Provide OAuth login for supported official providers. Discover CC Switch models from OpenAI-compatible gateways via `/v1/models` and `/models`.

by SteelUnion

DeepSeek Harness gate - per-turn persistent-operation counter: remind, instruct the model to stop, then ask for confirmation. Steps aside under sandbox/read-only modes; follows a live permission switch without restart. Verified on real hardware (full -> workspace-write -> read-only -> full). Crafted on an estimated 0.27 USD of model spend.