Plugins

Tracks API token usage and estimated costs with per-route model pricing, optional billing groups, synchronized CNY/USD display, session reports, composer totals and a clickable daily sidebar total.

Shows the active model's usage and remaining balance in the session header: OpenCode Go plan usage (rolling, weekly, monthly) with the selected model's monthly cap, Command Code 5-hour and weekly window usage with remaining credits, and DeepSeek account balance.

Cumulative token usage tracking for DeepSeek Harness, with per-model and per-day breakdowns, a settings-page dashboard and an online-changeable data directory.

AI-native expressions for two-way chat: people see images, while text-only models read fixed meanings; includes a picker, custom expressions, import/export, and two visual styles.

Session-local scheduled tasks: inject a prompt into the current session after a delay, at a fixed time, or on a repeating interval, so the model wakes up and runs it on time. Manage them from a panel above the composer, the /schedule command, or seven model-facing tools. Timers freeze while DSH is closed and never auto-resume; pause keeps the remaining time, so resuming continues from where it stopped.

Event-driven Hypatia memory for DSH: automatic conversation logging via session events, queued background consolidation (span summaries + work-unit extraction) with a dedicated model route, and recall injection — zero per-call model work in the main sessi

Maintained fork of dsh-approval-gate. A neutral operation whose confirmation count has reached the threshold auto-approves when the judge is unavailable, instead of prompting a human again; approver-facing explanations are generated in Chinese from the real sandbox mode, command and paths. Also carries deterministic hard-deny for credential exfiltration and system-path destruction, judge-input redaction, a judge model candidate chain, fingerprint-scoped allow rules, and an approval view with unified diff and one-click revert.

Wire-level LLM inspector: captures the raw HTTP request and response bodies of model provider calls and attributes each capture to the session that caused it through the llm/stream waterfall. The request pane renders the body as a collapsible tree with per-node sizes — one measured request was 98.6% a single 973-entry array — and the response pane reassembles the OpenAI-style SSE chunk stream into reasoning, content, tool calls and usage. Shown in a Conversation View tab and a loopback viewer page.

Rapid-refill circuit breaker for DeepSeek Harness automatic compaction. It refuses a compaction that would be futile — the context refilled within fewer than N tool turns M times in a row, which the unbounded step-pressure trigger cannot detect — and instead of burning a summarization call on every step it reports the situation with advice, usually that a single read or tool output was too large. Shows state and re-arms through /compaction-breaker, and injects one prompt section so the model relays it to the user. Covers host-plane sessions through its profile bundle; ordinary sessions need one row replaced inside an agent preset.

Sign in with a ChatGPT account and view Codex subscription usage directly in DSH Web's model settings, with browser or device-code OAuth.

Shows the account balance or coding-plan quota (5h/7d/30d) of the provider behind the current model, in one line under the composer; display only, no cost accounting.

Deterministic cross-session transcript retrieval for DeepSeek Harness: the model-facing `recall` tool searches past session logs with explicit scope control

Speaks your agent's reports in a voice you choose, compressing long reports before speaking. Character voices need no training - a 3-10 second reference clip clones one, and community-trained models work too - and no 6.4 GB GPT-SoVITS install: the plugin installs its own runtime and voice. An existing GPT-SoVITS can be used instead, and it is the same voice model.

Windows-native computer use for DeepSeek Harness: two tools drive and read any desktop app. One batch call runs find, click, type, key, read, wait and screenshot steps in a single engine pass, so a task whose steps are known up front costs one model round trip; reading returns window text rather than an image, and input and capture prefer UI Automation patterns, Win32 messages and offscreen PrintWindow so the user's focus and pointer stay untouched.

Use your DSH models in [Libre WebUI](https://github.com/libre-webui/libre-webui) for chat, sandboxed Work and usage tracking, with provider credentials kept in DSH. Requires Libre WebUI 0.37.0 or later.

Manual model probe for DeepSeek Harness: a session-view tab where you pick provider, API type, model and route, then send one real raw HTTP request over openai-completions, openai-responses or anthropic-messages — streaming or not — and read latency, TTFT, HTTP status, token usage and a three-layer diagnosis chain (API request / model response / strict check) on one screen, with errors in plain language.

Model-driven context compaction for DeepSeek Harness. The agent decides when and what to compress, commits durable session-log replacements, and ships reversible tool-result compression with byte-exact hash retrieval.

Registers the Xunfei Loomy cloud channel as an OpenAI-compatible supplier in dsh-router, adding model listing and streaming chat completions.

Shows the DeepSeek account balance as a row in the sidebar footer with a refresh button, and opens a popover with the balance card, daily token and estimated-cost charts over a 7/30-day window, a per-model table, and a link to the platform usage page. The balance comes from the official /user/balance endpoint; usage is aggregated from local DSH session logs and priced with the published peak/off-peak rate card in CNY, so it is labelled an estimate and its scope is stated on the panel.

Replaces the chat input model picker with a provider-grouped, collapsible list: groups models by provider, folds vision-bridge mirror providers back into their upstream, and carries the reasoning-effort options each model declares in settings. Tested on DSH Desktop only; the DSH web build is untested.

A full-page console panel opened from the sidebar, with eight tabs: an overview; a session list read from the sessions service (title, directory, running state, updated time, with filters); installed bundles and Loader entries from pluginInventory and pluginManager, with per-bundle enable switches; a plugin market listing public topic:dsh-plugin repositories from the GitHub search API; a model and provider editor that writes the llm-pi-ai settings namespace through remote.settings.mutate (switch a session model, add or edit providers and their model lists, and store API keys via remote.credentials, merging over the stored profile so fields the form does not expose survive an edit); MCP and skills tabs; and a usage tab that reads the host's loopback /api/dsh-usage/overview document. Services are declared through ctx.inject, so a deployment that omits one mounts without the console rather than crashing.

Expose every model already configured in your local DeepSeek Harness to the LAN over OpenAI- and Anthropic-compatible endpoints, with API keys, per-key rate limits, a circuit breaker, model filters and usage reports.

AlphaRSI — three-role self-improvement loop that evolves the training recipe of alpha sequence models. Operator runs walk-forward, Evaluator grades with OOS RankIC, Refiner proposes hyperparameters via TPE Bayesian search; only strictly-better recipes are kept. Offline one-command run, data bundle shipped with sha1 reconciliation, GPU auto-detect. Ships one skill.

Adds configurable HTTP header profiles that can be selected per model.