Plugins

Five tools for reading a video as raw material rather than as something to watch. video_outline maps a whole file as text and draws nothing at all: 4-60 equal windows, each with its cut count, speech share, intensity bar and tags, then the windows worth opening and the exact start/end for the call that follows -- zero image tokens. video_probe returns metadata and a reading plan without decoding frames. video_storyboard turns a window into one contact sheet plus a text index binding every cell to its timecode, cut count and frame time; cells whose text would be illegible at that scale are marked zoom, and dedupe global folds a picture that comes back after a cut. video_zoom re-renders 1-9 cells from a short window at a larger scale. video_read decodes only keyframes, OCRs each screen, and returns an outline plus \[mm:ss] text -- the path for lectures, slide decks and screen recordings, where the words are the content. It also reads a file with sound and no picture as a transcript (subtitle track, then DSH_ASR_CMD, then a transcript file you pass; long audio is cut at 45 s onto silence), composes six changed frames into one sheet when a file has no readable text at all, and can write the result out as SRT, WebVTT or text. Cuts, dissolves and one-frame flashes are found in a single ffmpeg pass; from ten minutes up a file is analysed on its keyframes only, 2.7x faster on a two-hour lecture; and a cell is never shrunk below a fifth of the source, so a 1440p sheet spends the same tokens on fewer, bigger cells. OCR is the engine Windows ships, tesseract with line boxes, or DSH_OCR_CMD. It takes a B station link (bilibili.com/video/BV..., a bare BV id, an av id or a b23.tv short link) wherever it takes a file: up to 300 seconds is read from the CDN in place with nothing written to disk, longer videos are downloaded once and cached within a 4 GB / 24 video budget, play URLs are cached for ten minutes, and a link that is not a single video says what it is. Needs ffmpeg, uses ffprobe when present, and has no other runtime dependencies. It is a wire-compatible reimplementation of SAOS-3624/dsh-video-storyboard: it registers the same video_probe / video_storyboard / video_zoom tool names, so the two cannot be enabled together.

Bridges the Codebase Memory MCP code knowledge graph into DSH and keeps indexed repositories fresh with a debounced filesystem watcher. Windows-only MCP bridge (the watcher half is cross-platform); auto-installs its npm CLI via npx on first start.

Search, list and read past and current sessions through the harness's own `sessionQuery` service: uses the SQLite FTS5 index where a deployment enables it, and falls back to a bounded newest-first scan where it does not.

Lazy capability catalog for DeepSeek Harness: one compact index plus search/describe/call meta-tools for MCP tools and skills, instead of every schema on every request

A full-screen novel-writing workbench: chapter board, draft candidates reviewed and merged paragraph by paragraph, plot summaries carrying an index header, and a writing pack assembled from the outline, character cards and earlier chapters. It also reads a shelf of finished books you import, extracts their writing habits in one run, merges them into a single style standard, and puts that standard in the prompt for every later generation.

Discovers rule files from every convention OMP supports and injects them into the system prompt as an always-apply layer plus a rulebook index, then enforces them with time-traveling stream rules that abort violating output mid-turn, deny a matching tool call, fold a reminder into that call's result, or ask a judge model about the finished reply. The same tool lets the model write new rules, and an audit panel on the Plugins page says why each rule is in force.

Cache-first code exploration: the rdd_explore tool queries the project's .rdd/exploration index and, on a miss, dispatches a read-only worker subagent to explore and register findings; rdd_models serves the live model catalog for dispatch-time model selection.

Turns a local Obsidian vault into a dedicated RAG chat panel: a sidebar page with the note browser, the retrieval-augmented chat and the cited source side by side, answering from fully local BM25 plus CJK-bigram retrieval over your notes and streaming through the DSH model channel, with clickable \[n] citations and mtime-based incremental indexing.

Shows which graft index a dsh session is serving and whether it still matches the code, and gives the model graft_ask, graft_grep, graft_callers, graft_skeleton and graft_map as native tools.

Model picker for the DSH web composer that replaces the model seat: a search-first popup whose left column indexes the suppliers and scrolls the grouped list, with a favorites group pinned at the top, image icons separating native from bridged vision input, and a reasoning-effort panel opened from the right zone of the trigger.

DeepSeek Harness Web plugin — PDF text-layer and scanned-PDF OCR, Office docx/xlsx/pptx to Markdown, TIFF/epub and long-document index cards.

Codex-style `@file` mentions for the web GUI: search and reference workspace files of every format (.java, .vue, PDF, images, extension-less, hidden) with no index truncation. Replaces dsh-at-file.

Offline mirror of the official DeepSeek Harness documentation as a dsh skill: 168 site pages (EN/zh) plus 52 repository-only reference docs, with topic navigation and a full-text search index for plugin development.

HarnMem memory plugin for the DSH (DeepSeek Harness) agent: byte-stable hot-index injection, memory tools, session-end extraction, and a read-only memory browser panel.

Folds oversized tool results into on-disk spill files with deterministic head and tail markers, indexes compaction summaries in a searchable ledger, and extracts cross-session memories with per-project scoping. Any folded content is restored byte-for-byte on demand through one tool call.

Layered long-term memory for DeepSeek Harness: each turn injects a persona frame plus a relevance-ranked memory index while full transcripts stay on disk and are expanded on demand, so the vault never floods the context window. Facts are slot-keyed, so a changed mind retires the old value instead of recalling two contradictory answers. The persona frame is derived from memory by the model itself, behind three anti-pollution gates (rejecting development chatter, blocking echoes of its own injection, and merging restatements by shared content words). Periodic consolidation merges near-duplicates, rolls long topics into digests, and archives low-value old entries by tagging only — raw text is never deleted. The append-only store is compacted in place (measured 4.7 MB to 355 KB). Zero third-party dependencies, no build step. Note: per-turn injection inevitably lowers prompt-cache hit rates, so this plugin is not recommended for work scenarios.

Cross-harness agent memory for DSH: one plain-file store shared by the DSH and pi adapters and the harnmem CLI, so a memory written in one harness is available in another; injected as a byte-stable hot index, with five harnmem_* tools, session-end extraction, and a read-only memory browser panel.

Structural long-term memory for DSH: supersede, decay and entity-state tracking across 15 MCP tools, automatic recall and capture hooks, and a bounded pointer index published per request instead of filling the window. Requires Python 3.10+ with torch and sentence-transformers; the first run downloads a ~470 MB embedding model.

Skill index browser with search, custom alias editing, and chat-input skill insertion.

Continues work across coding agents inside DeepSeek Harness: six sessionflow_* tools query a local session index built by sessionFlow to find past work, show where the current project left off, and hand the goal, files and open items to the next agent. Requires the separate sessionFlow core, installed from its GitHub repo. Everything stays local — no cloud, no telemetry.

AOCI cognition panel for DSH: multi-workspace AOCI overview, per-repo index and governance view, MCP binding diagnosis, and guided or batch initialization. Governance, verification and guided/batch initialization require the external aoci CLI on the host; index browsing works without it.

Long-term memory for DSH stored as plain Markdown. Real user/assistant exchanges are captured into daily session files without a model call; an LLM pass then distils them into memories that are injected into the system prompt each turn. Retrieval is local (CJK bigram + BM25) in three levels — index, summary, full text — so the model only pays for the detail it asks for. Memories carry priorities; P0 and user-written entries are never changed automatically, and forgetting marks an entry stale instead of deleting it, so the whole chain stays auditable in status.md. Six tools cover recall, write, status, extract, feedback and the skill library; when a cluster of memories is ready to become a skill, a card appears at the end of the turn, writes the SKILL.md on click, and installs it only after you approve. Learning is on by default and follows the conversation model; turning the master switch off stops the LLM entirely while capture keeps working. Failures are surfaced in the prompt and in status.md instead of being swallowed. Pure JavaScript, no third-party runtime dependencies beyond the harness packages.

Plugin index inside DSH that lists over 10,000 third-party plugins checked against the dsh.bundle manifest in their package.json rather than by star count, with one-click install and a companion website.

Study coach for DeepSeek Harness: turns a course into a knowledge map of subject, module and unit, records mastery per unit in six stages, plans daily tasks against a materials index, keeps a wrong-answer book and spaced-repetition cards, and serves its own web panel.