Plugins

Extends dsh-subagent-tools with a per-call cwd for subagents, shipped with the two in-process provider patches it requires.

Per-call model, provider, persona, and toolFilter overrides for subagent delegation, with @preset: references and provider/model composite ids.

Update manager for installed plugins: check for updates, back up, and roll back.

Plugin center: discover, install, and manage DSH plugins from Settings.

Per-agent on-demand tool discovery and progressive schema disclosure.

Project .pack.json/.pack.zip into .agent-pack/modpacks/ and expose skills via a workspace allow-list.

Bring your existing Agent Skills (SKILL.md) library to DSH: discover skills across Claude/Codex/Cursor/Gemini paths, inject a progressive-disclosure index, and load bodies on demand.

Bounded, cell-addressed `doc_read`/`doc_write` for xlsx / pdf / docx / pptx / ipynb, plus an MCP server and CLI.

Build auditable KB packs (SQLite FTS5) from md/txt/docx/pdf with deterministic retrieval and original-text reading.

Codex-backed `web_search`, `image_gen`, and `image_vision` tools for DeepSeek Harness, reusing ChatGPT OAuth login state.

Native LLM-provider vision bridge: images pasted in the chat are described by a vision model (Qwen3-VL via pi-ai/llama.cpp) and the text description is fed to text-only DeepSeek for the reply — image admission, routing and compaction all run through harness-native mechanisms, with an LRU description cache and 503 retry.

Composer-attached images are transcribed to text by an OpenAI-compatible vision model before reaching text-only DeepSeek models.

One shared Obsidian vault for every agent: dependency-free Python core (search/promote/adjudicate/forget), vault template, dsh plugin (memory_search/show/submit/status).

Project-scoped cross-session memory: PROJECT.md snapshot injected into the first user message, a memory_remember tool, and auto-reflection after ReAct tasks; each project keeps its own memory file.

Bounded local memory with CBDC authority gating: SQLite + FTS5 claims, scoped recall with explainable use/verify/ignore decisions and a full audit trail, /memory commands, ≤3-claim/1200-char injection per call, no extra model call.

Cited memory over DSH's lossless session log: distilled facts carry `(sessionId, eventRange)` citations that expand back to the exact original log excerpt.

Render local image paths from assistant replies inline in the message body (9 formats, click-to-zoom lightbox, adjustable size).

Copy and open `dsh://` session deep links, or paste them into another conversation to inject a bounded, read-only snapshot of the referenced session.

Edit user and built-in system-prompt sections with live preview.

Cross-session task queue with handoff notes: sessions and subagents push, claim, complete and cancel tasks on a shared file-backed queue, and leave a handoff summary for whoever picks up next.

Cross-session messaging for DSH: any session on the machine can list and message any other, Claude Code-style, via a local heartbeat registry and inbox.

Font switcher for the DSH Web GUI: 99 UI fonts and 31 code fonts with CJK-Latin pairing stacks, instant apply and localStorage persistence.

Windows tray desktop shell for DSH: auto-installs the exe from GitHub Releases, creates a desktop shortcut, and adds a desktop_launch tool to start it from chat.

Activity heatmap in the DSH Web sidebar: GitHub-style grid of daily commits, token usage, and estimated spend, with a today stats line for all-session token totals, cache hit rate, and per-model auto-priced cost.