Plugins

Keeps one reasoning-only turn from making an entire session unusable. When a turn produces no visible text and no tool call, the assistant message is persisted empty, the gateway then rejects every later request in that session with 'content or tool_calls must be set', and the session is dead with no way back through chat. This plugin registers a single llm/stream waterfall listener and, only when the turn carried nothing visible, injects a small text block immediately before the terminal finish, so the persisted message is never empty. Ships the fixture DSH's own mock cannot express (a strictly reasoning-only SSE server, since llm-mock-server forbids an empty successText and always appends text after reasoning), plus an end-to-end reproduction that replays a real persisted session through DSH's own serializeMessages. Preventive, not a repair. Zero dependencies, one file, no process or filesystem access.

A calmer reading view for DeepSeek Harness: native process details, live reasoning, and clean final answers.

Model channel manager: several provider/model candidates share one virtual route (roundrobin/<group>) with failover, retry, timeout and cooldown; a settings tab edits providers and groups, runs per-model speed tests and real single-model request tests, and a seven-day health panel aggregates latency, failures and upstream token usage (input, output, cache and reasoning).

Declares per-model reasoning-effort ladders for third-party providers in the llm-pi-ai settings section, so DSH's stock thinking-intensity selector works on models the pi-ai catalog does not describe.

ChatGPT-style side chat for the DSH Web GUI: ask small questions about the current topic in a floating side panel, answered by an independent lightweight session over the official /api RPC — the main conversation is never touched. Streaming text, collapsible reasoning, tool chips, stop button, agent-preset selector, and an automatic context snapshot from the visible main chat.

Session model selector with search: two-stage loose matching (normalised substring first, subsequence as fallback, so glm53 finds GLM-5.3 and ds finds DeepSeek), providers with recent successful calls pinned first, and a reasoning-effort pane that remembers each model's last explicitly chosen effort.

Adds per-route context window and compaction threshold fields to the Models settings page and a vendor-aware thinking-effort picker to the composer, writing llm-pi-ai capacities, reasoningEfforts, and a compaction-basic patch row.

Five web UI tools in one plugin: a provider + model two-button selector with reasoning-level control, collapse/expand-all for sidebar workspace groups, a session tab listing the files the session changed (needs kernel 0.1.2-alpha.1+, silently absent on older kernels), a workspace badge next to the session title, and a settings page that toggles them.

LiveBench leaderboard for DSH — discovers the release the official site currently publishes, then reads its CSV and category JSON, and renders model rankings in an isolated same-origin iframe opened from a button beside the composer. Search, organization / open-weight / reasoning filters, per-category sorting, task-level detail, up to four models side by side, CSV export; refreshes every 60 s by default (configurable 30–3600) and caches only to the plugin's own .cache/rankings.json.

Read-only session work-path observer: fold turns into stage, run, duration and tok/s summaries, then expand parallel branches, tool results and vendor-visible reasoning.

Measures what an llm-pi-ai endpoint really accepts — context window from its own model list, output-token limit and reasoning levels read out of its own range errors, image support from a generated test image — then reports declared values that contradict that evidence and rewrites them only after confirmation.

Settings section that reports token usage: totals for input, output, cache read, cache write, reasoning and call count, with breakdowns by model, day, month and project plus a per-day trend chart for the last 30 days; history is rebuilt once on a worker thread and then kept current from live session events.

Replaces the native DSH reasoning-level option list in place with a draggable Off, Low, High, and Max slider while preserving the official model-selection flow.

DSH (DeepSeek Harness) 插件：为每个 LLM 模型配置默认推理等级（reasoningEffort），并在请求进入底层调用链前按路由规则自动补齐

Injects one system-prompt section so the agent's chain-of-thought, planning and tool-invocation reasoning stay in Simplified Chinese. Zero external dependencies - clone and link-install from any local directory, no npm build or publish required.

Dual-endpoint Zhipu BigModel GLM provider catalog (Coding Plan and ordinary API) with live model discovery, live-tested thinking-tier mapping, and a settings card for keys, endpoints and the default reasoning tier.

Tencent CodeBuddy model provider for DeepSeek Harness that signs in with browser OAuth instead of an API key, lists the CodeBuddy model catalog with the context, output, tool-calling, reasoning and image capabilities of each model, keeps multiple accounts with automatic failover to the next usable one, shows remaining quota in the composer with a token and credit usage panel, and completes automatable CodeBuddy growth tasks per account with an execution log. Install from npm with `dsh plugin --profile web add @tnnevol/dsh-codebuddy`.

Auto-fills reasoning-efforts, context windows, output limits and image modality fields for models from unofficial (custom) providers (with data from models.dev), and provides user experience improvements such as compatibility enhancement and remembering reasoning effort.

Grouped provider and manufacturer model selector for DeepSeek Harness, preserving provider/model routes and reasoning levels.

Prompt enhancer for the DSH web UI with basic/standard/expert modes: standard/expert modes rewrite drafts with project context and recent session history, retry gracefully when gateways reject reasoning-effort settings, and show a before/after review with undo.

Pins every request to the native max-effort tier and adds a per-model/per-mode reasoning-depth panel with a 0-20 capability slider, isolated Velocity/Apex profiles, and offline reasoning toolboxes. Distributed as an obfuscated build; commercial use allowed, reverse engineering closed.

Context-window attribution for DSH Web: a Mini Utility Dock panel that ranks the user messages, assistant turns, reasoning blocks and tool results of a session by the tokens each one occupies, so it is clear which entries hold the window. It also shows the share by entry type, the heaviest entries across the live sessions of the host, and, where the host provides them, provider usage and the estimated system/tools/messages composition. The panel is read-only: it does not estimate cost, modify a session, or trigger compaction.

Auto-adjust model reasoning effort per task (off/low/high/max): task classification, \[effort=...] markers, in-turn error escalation, DeepSeek peak-pricing-period capping with countdown notices, and manual-selection priority.

Maintains reusable named subagent profiles for DeepSeek Harness and runs them with per-profile model, reasoning effort, and agent-preset settings.