Plugins

Security-analysis plugin for DeepSeek Harness: 33 tools spanning six security domains, evidence/reporting and toolbox, plus the case lifecycle, routing and tool ledger; a session-bound case workflow with an evidence chain, E-id validation and reproducible findings; 362 on-demand references; three hot-swappable persona profiles; and a four-tab workbench (armor / case / intercept / about) with case-state recovery after context compaction.

Token optimization for DeepSeek Harness with searchable MCP tool discovery and lazy loading, configurable compaction controls and status reporting, tool-output reduction, long-text vision summarization, and opt-in conversation archiving with paginated retrieval.

Use a ChatGPT/Codex subscription in DeepSeek Harness with GPT models, image generation, web search, subscription quota reporting, model and feature controls, and browser or device-code OAuth sign-in.

Layered token-optimization pipeline for DeepSeek Harness: output ladder, MCP lazy loading,compaction driver, cache-hit reporting. Built on real DSH plugin APIs; ~40-60% input saved in long sessions.

Kanban task board for DeepSeek Harness sessions: the main agent dispatches tasks to one-shot worker/verifier subagents, with dependency scheduling, file-level exclusion locks, arbitration hold/freeze, milestone progress reporting, and per-task token usage stats.

Evidence-driven goal verification plugin for DeepSeek Harness. Locks goals via /ultragoal and verifies completion against command exit codes and generated artifacts, preventing the model from falsely reporting task completion.

DeepSeek Harness plugin: makes a reverse-proxied (non-loopback) dsh web page count as a trusted host, so the Settings surface loads instead of reporting "settings are unavailable in this browser".

Local images and videos as the DSH desktop window background: timed rotation with 0.7s cross-fades, first-frame-ready before fading in, automatic light/dark ink for readability, adjustable blur and panel/sidebar opacity, video thumbnails, and a probe endpoint reporting load, layer and selector health. Reads only the local files you point it at.

Tracks DSH's file reads, image views, and code edits in WakaTime, including exact counts of lines added and deleted by AI. Keeps activity tracking running while DSH responds, batches updates across files, and coordinates per-project reporting limits across parallel DSH processes. Automatically installs and updates wakatime-cli.

Registers a wrapper provider that gives delegated subagents their own LLM route — one-off spawns through provider.start, Agent Teams members through a hook on startActivation — while the main agent keeps its model. Ships a Settings section for picking that route, a per-expert override file, a capability flag reporting which seams the hook caught, and a readiness record written at mount.

Adds a Delete session action to the sidebar context menu behind a confirmation modal reporting log size and live activity, cleanly unlinking log files, projection checkpoints, and host memory registrations.

Connect several OpenAI-compatible gateways at once: each one is its own group in the model picker, can be hidden without losing its settings, and can fetch its own public info from its address, alongside per-gateway model management, model tests and optional quota reporting.

Diagnoses third-party LLM API endpoints from the sidebar: a four-layer check (DNS/TCP/TLS, key and model, a real streaming conversation, then timed repeat runs) reporting time-to-first-token, output speed and success rate, with every failure attributed to the step it stalled at and the byte count that actually left the machine. 37 provider presets, no model names hardcoded.

Econometrics research assistant: 6 tools covering method selection, data preparation, model specification, empirical analysis with Python/R/Stata code templates, robustness checks, and result reporting.

Organize DeepSeek Harness workspace sessions by renaming them as 类别｜主题 and reporting archive, rename, and review suggestions.

Registers one tool, report_health, that answers whether a scheduled report job actually ran. It judges by side effects rather than exit codes: how old the newest report folder is, how many documents it holds, whether more than one report category is present, whether the audit ledger was synced recently, and how long ago the last log was written. Exits nothing, calls no model, and reads only local paths you can override by config or environment variable.

Read-only panel reporting an instance's plugin bundles, patch inserts, skills, upgrade health and decoupling health.

Lets the model trigger compaction itself at a safe point via a compact_now tool (works mid-turn), plus a context_status tool reporting real token usage against the session's actual context window.

Keeps an unfinished task list visible across turns. DSH clears the todos projection at the start of every turn and expects the model to rewrite the whole list, so when a turn is interrupted the plan is durably in the session log but permanently gone from the UI — and todo_write is write-only, so neither the user nor the model can read it back. This plugin registers one ordinary session/event listener, remembers the last list per session, and when a turn/start arrives whose remembered list still has unfinished work, re-appends that list so the projection stops reporting an empty plan. No fork of tool-todo and no waterfall: a plugin cannot replace the projection (the registry throws on a stateVersion mismatch), so appending the event the fold reacts to is the reachable seam. Deliberately narrow — only unfinished lists, at most one event per turn start, fails open. Verified by a controlled A/B on real sessions: identical runs differing only in whether the plugin was installed, folded through the real projection definition. Zero dependencies.

DeepSeek-Harness plugin for multi-model orchestration with recursive task decomposition, dynamic routing, quality gates, failure recovery, and KPI reporting.

Deterministic review of synthetic AI-agent claims in DeepSeek Harness, reporting bounded verdicts, reasons, and evidence references.

One-step local model onboarding plus a light harness for DSH. On load it probes loopback and private-network inference endpoints (llama.cpp, Ollama, vLLM, LM Studio), reads each model's context window and output budget from the backend, and registers the provider through settings, so a local model becomes selectable without hand-writing config. It then trims only that provider: masks the tool list through tools.restrict, drops named system-prompt sections and memory contexts, swaps the injected instruction block for a short one while preserving its first 80 and last 96 characters, and truncates tool-schema descriptions without touching structure or required fields. Cloud providers are returned as the same assembly object, untouched. The injected block is re-read from disk on every assembly, so editing the file takes effect on the next request with no reload. The panel in the plugin detail page is a live operations view: it shows whether the takeover is active and by which hook, how many characters the last assembly was trimmed from and to, and which providers are online (probed concurrently, with a real one-shot completion reporting latency and reply text per provider). If a local backend is not running, the panel offers a one-click launch: it probes the port first (never starting a duplicate), then starts the matching server (ollama serve, or a custom command you supply) detached and polls until the endpoint answers, clamping the wait to 3-600 seconds, reaping the child it just spawned if it never answers, and naming the paths it searched when it cannot find the binary. A stop button kills backends this plugin started (SIGTERM, then SIGKILL) and refuses to touch a server somebody else launched. When a model has just been downloaded and no backend exists yet, the panel can also scan the machine for .gguf files with hard caps (300 models, 20000 directory entries, Downloads only one level deep), start llama-server on the chosen one (skipping the .mmproj. vision projection and reusing a server that already answers), register the provider under the model id the backend reports, and run one real completion to prove it works. Before spawning anything it compares the model file size against available memory and refuses a model that cannot fit (or that exceeds 80% of the budget) until you confirm, switching to --no-mmap in that case; the honest limit is that this reads OS-visible RAM only, not GPU-side memory.

Checks enabled plugins for newer versions at startup, lists them in a corner card and upgrades them all with one click, reporting per-plugin success or failure; resolves the bundled dsh CLI on macOS, Windows and Linux instead of assuming a Windows executable.

Step-by-step workflow map for DeepSeek Harness: folds a session into one step list, labels where each percentage came from rather than reporting a bare number, and queues a prompt for the next step with revoke before it is claimed.