Plugins
Browse, filter, and install DeepSeek-Harness plugins.
99 plugins found
dsh-deepcanary
oscar-williams/dsh-deepcanary
Evidence-first attention supervision for DeepSeek Harness with C0–C3 policy, deduplicated Inbox alerts, quiet hours, and redacted outcome records.
dsh-context-doctor
zhenyu98/dsh-context-doctor
Context injection audit: token costs of instruction chains / skill catalogs / tool schemas, duplicate and conflict detection.
dsh-memoir
qinling-melon-farmers/dsh-memoir
Local-only project memory with zero bundled runtime dependencies or external services: JSON as the source of truth plus generated PROJECT_MEMORY.md, token-bounded session-frozen Hot Memory for prompt-prefix cache stability, cached cross-project BM25 recall, trusted session/turn provenance, explainable pre-write duplicate/conflict governance, editable importance/pin/tags/supersede/archive lifecycle, configurable top-level turn-end distillation, and a bilingual Web GUI for search, inspection, diagnostics, and live settings.
dsh-edit-diff
better-er/dsh-edit-diff
Client-side re-render of edit/write tool cards with a near-linear line-level diff that drops duplicate unchanged lines and underlines the actual changed characters inside replaced lines.
dsh-localtts
abbccdd/dsh-localtts
Reads Harness replies aloud through an existing local TTS runtime: built-in IndexTTS and GPT-SoVITS launchers, a protocol-based local HTTP/JSONL entry, and model-folder loading for Kokoro, F5-TTS, VoxCPM, Qwen3-TTS and CosyVoice SDK environments. Ships no models. Derived from 1624318455/dsh-plugin-tts; it isolates its row id and settings, so both can be installed side by side with duplicate features deactivated.
dsh-win-pwsh7
xinspark/dsh-win-pwsh7
Reports which PowerShell the DSH pwsh executor actually runs (enabled rows only; version-numbered Store paths are never recommended as pins), and installs a Store-managed PowerShell 7 through a scripted Store-page click or the manual Store path. Allows the pwsh tool only when the live executor resolves PowerShell 7+, and when dsh-ops is installed keeps the native pwsh beside ops_bash by enabling Compatibility and hiding duplicate official tools. Native status card and configuration switch in the Web UI; Windows only, no runtime dependencies.
dsh-skills-anywhere
noteflowai/dsh-skills-anywhere
Live skill provider that adds skills from 60+ other agents' directories (Claude Code, Codex, Cursor, Gemini CLI, Copilot), Claude Code plugin marketplaces, and any git repository of skills to the dsh catalog without copying or symlinking; deduplicates, repairs frontmatter, renames name collisions, and ships a CLI (list, add, sync, doctor).
dsh-biomemory
klrsl/dsh-biomemory
Biomimetic memory system for DeepSeek Harness: a single SQLite store (node:sqlite, WAL) whose Markdown tree is kept as a read-only mirror; memory tools, frozen-snapshot injection at session start, graded approval gating, on-demand extraction of the current session, near-duplicate merging, an archive and supersede lifecycle, memory metabolism (dream), memory pins, semantic recall over local embeddings, and the /memory command.
dsh-session-explorer
zn-dk/dsh-session-explorer
Message-level full-text search across DSH sessions via an FTS5 trigram index (user / assistant / steering / tool with per-kind filters), deduplicated results across fork/continued sessions, read-only context preview with auto-scroll to the focused message, one-click jump to the real session, and incremental/full index rebuild with health check; zh/en i18n follows the Host locale service.
dsh-pubmed
aiyacharley/dsh-pubmed
Literature research toolset registering 26 model tools over PubMed, Europe PMC, OpenAlex, PubTator3 and Semantic Scholar: field-syntax search, ID conversion, bibliographic lookup, citation formatting, MeSH and related-article walking, PubTator bioconcept resolution and relation queries, Europe PMC and Semantic Scholar paging, full-text retrieval, and open-access PDF discovery with optional download. A unified search deduplicates hits across sources and orders results by cross-source hits, then relevance, then citation count; a session knowledge graph with PubTator concept enrichment commits to a persistent user graph and renders as Mermaid. Most sources need no key.
dsh-skill-mover
mjylfz/dsh-skill-mover
Migrates skills from other agents (Cursor, Claude Code, Codex, OpenCode, and more) into DSH: scans 14 agent platforms plus the shared ~/.agents skills layer, detects duplicate names across sources, lets you pick the best source per skill, copies them into the DSH skills directory, and rolls back previously migrated skills.
dsh-bili-interactor
acttk/dsh-bili-interactor
Browses Bilibili on your behalf and actually takes part: search videos, like, coin, send danmaku, post comments, reply to other people, and keep replying when someone answers back. Reads the video's Bilibili AI subtitle (or transcribes the audio locally when there is none) so its comments talk about what the video really says, and it reads that same content before choosing whom to reply to. Instead of copying other people's sentences, it distills rules about how to sound like a person and gets less AI-ish the more it runs. Skips videos it already browsed, de-duplicates replies per thread, and after too many rounds with the same person in one comment section it reviews whether the exchange has any substance -- if it looks like two bots complimenting each other, it says one last line and stops replying to that person there. Checks whether each comment is genuinely public or only visible to your own account, can scan any time window (last hour / day / 3 days / all) and lists the ones others cannot see, warning you on the overview when replies got folded. Six-tab settings panel where every change takes effect immediately with no restart, a manual run that queues behind the background listener instead of being refused, and an environment self-check that tells you exactly what to install. Requires Python 3 for the Bilibili actions; the built-in check lists any missing packages.
dsh-fix-duplicate-loader-id
zhao1012/dsh-fix-duplicate-loader-id
Diagnoses and fixes "duplicate loader entry id" boot crashes - finds duplicate loader rows across profile bundles and converts them into id-targeted patches via a zero-dependency CLI (detect / preview / fix / verify).
dsh-mnemosyne
i11ustrious/dsh-mnemosyne
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.
gitee-ai-employee
wangbobo-coder/gitee-ai-employee
Gitee/GitHub AI employee: @ the bot on an issue (Gitee or GitHub), it clones the repo, implements the fix on the branch you name, opens a PR to that branch, and can auto-merge and auto-close the issue. v1.2 adds security scanning: configure scan repos, audit with built-in/custom prompts, deduplicate, and file new findings as issues.
dsh-skill-panel
ztqing4agents/dsh-skill-panel
Panel for the DSH web GUI that keeps built-in, local and market-installed skills in one list, decides each row by on-disk evidence (symlink, install ledger) under a fixed precedence, listing a duplicated name once, from the higher-precedence source. Switches rewrite disable-model-invocation in its own SKILL.md; built-in skills have no writable file, so the panel stores their state and says the registry is unchanged; symlinks are read-only. The add dialog takes a local directory or a public github.com link (.git, /tree, #path:), shallow-cloned. A market tab browses the skillhub.cn mirror across three source scopes and installs into ~/.dsh/skills, with update checks and rollback on failure. UI ships in Chinese and English. Package name is @ztqing4agents/dsh-skill-panel (npm) and the mount id stays dsh-skill-panel, so it installs alongside dsh-skill-center. Its HTTP routes have no DSH access check, so a web page open in your browser could call them.
dsh-local-trim
leonlee114514/dsh-local-trim
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 does a fit check that counts only the part of the file the GPU cannot hold: hostBytes = fileSize - VRAM, compared against max(total RAM - 2GB, free RAM). A file far larger than RAM is refused outright; a heavy host-side remainder is refused unless you confirm; the same remainder under mmap is only a warning, because the weights land in reclaimable page cache. A load-mode selector (auto / mmap / none / mlock / mmap+mlock) maps to llama.cpp --load-mode, and the default is whatever the engine picks. From 0.4.2 everything above is derived from the platform instead of assumed: home and data directories come from os.homedir() and the OS conventions (LOCALAPPDATA on Windows, ~/Library/Application Support on macOS, XDG_DATA_HOME elsewhere); backend binaries are looked up in the official install locations of each OS and can be pinned with OLLAMA_BIN, LLAMA_SERVER_BIN, VLLM_BIN or LMSTUDIO_BIN; model discovery walks DSH_LOCAL_MODELS_DIR, ~/models, ~/llama.cpp/models, ~/.cache/llama.cpp and the LM Studio data directory, with the WorkBuddy layout demoted to one candidate among them. VRAM is asked from nvidia-smi, then rocm-smi, then sysctl hw.memsize on Apple Silicon (flagged unified, so the same bytes are not charged twice), and any flip to a unified-memory budget is handled explicitly. Before building a command line the plugin runs llama-server --help once and keeps only the flags that build actually has, so older engines get --n-gpu-layers 99 and no KV quantization while newer ones get auto plus q8_0; anything it cannot probe is passed through rather than dropped. LM Studio is started with lms server start --port N rather than being lumped in with the engines that need a model argument, and vLLM gets its own --model example instead of llama.cpp syntax. Windows, macOS and Linux are all exercised in CI, including a windows-latest and a macos-latest job, because a cross-platform claim that has only ever run on one OS is not a claim. Two of those suites exist specifically to catch this machine's blind spots: one replays the whole suite with process.platform rewritten to each OS, and one deletes every advantage this development box has — no nvidia-smi, no rocm-smi, no sysctl, an empty PATH, a home directory with no models — and requires graceful degradation rather than a crash. That last one caught a real class of bug: several assertions hardcoded 'lms.exe', which cannot pass where the executable has no suffix. Measured on the reference machine: a 16.9GB 27.2B MoE with 14.5GB of VRAM holds a 285-691MB working set and sustains 76 tok/s after a 1000-2000 token warm-up, so file size alone is not a fit signal. From 0.4.3 the scanned models are shown as a table rather than a dropdown (one row per gguf, with its size and any fit warning), so you can see what is on disk before choosing. From 0.4.4 the delete button on the provider table removes the provider entry, not the weights file on disk. That distinction matters: what people actually want to clear out is the dead entry - one pointing at a port that never answered, one registered twice for the same port, one whose model file was deleted long ago - not a 15.8GB gguf. Deleting a config is undone by importing again; deleting a gguf is not. The removal goes through settings.mutate with an unset op on \['providers', id], because update merges layers: writing null there leaves a null provider key that the loader reads as a broken config rather than an absent entry, and only mutate/unset is implemented as Reflect.deleteProperty. The result is read back afterwards, because unset on a key that comes from a lower layer (the profile cordis.patch.yml or a home patch) does not delete it - it restores the inherited value, so the call succeeds while the entry stays. When the entry is still there the panel says so and points at the file to edit instead of reporting success, and a DSH whose settings service predates mutate() is told plainly rather than silently ignored. Deleting takes two clicks and clearing the confirmation is as easy as clicking another row; on success the online and baseURL caches are cleared and the tool surface is recomputed, since the provider list changed. The gguf itself is never touched by this plugin - the scanned-model table is display and selection only.VRAM reporting is driver-dependent. From 0.4.5 model discovery also reads Ollama's own storage: its blobs are named sha256-… with no extension, so extension matching could never list them. Weights are now accepted by extension OR by GGUF magic header (only dot-free files of at least 8MB, capped at 600 header reads per scan; obviously-not-weights extensions like .json/.md/.safetensors/.part are never opened). Ollama models are resolved from their manifests: the model layer is picked by mediaType application/vnd.ollama.image.model (never adapter/mmproj layers), the human-readable name is restored from the manifest path (…/qwen3/8b becomes qwen3:8b, …/latest stays bare), tags sharing one blob merge into a single row, and the metadata kv supplies name, quantization and context length. Adopting an Ollama model does not start a duplicate llama-server over the same weights: it confirms ollama serve answers on 11434 (launching it if not) and registers the provider under the tag the endpoint itself reports — the blob's sha256 name would never match anything, since 11434 lists qwen3:14b, not sha256-a8cc…. Two real-machine-only traps fixed on the way: existsSync('<file>/') returns true on Windows, so manifest-vs-directory branching now goes by parseable content instead of trailing-slash semantics; and the quantization label table had been copied from a stale ggml enum generation, labelling ggml_type 15 as Q5_K_S while ollama show reports Q4_K_M. Same-blob dedup is by (dev,ino), keeping whichever entry has a name. Verified on the reference machine: the scan goes from 1 row to 4 — qwen3:14b (8.64G) and qwen3:8b (4.87G) now appear with Q4_K_M badges — in 4ms; npm run verify exits 0 (launcher 57, portable 35, panel-host green including the linux/darwin simulated-platform runs); the new panel-host suite builds a real miniature Ollama store in tmpdir and points OLLAMA_MODELS at it, so results no longer depend on whether the test machine happens to have Ollama installed. From 0.4.6 the panel shows each model its real name instead of its file name, because some download tooling stores every weight under one generic file name, leaving the real name only inside the GGUF header. The scan now reads each file GGUF KV section - not a fixed-size head buffer, because tokenizer.ggml.tokens (hundreds of thousands of strings on a large MoE model) and the chat template push general.file_type past any window; and GGUF string arrays are interleaved (u64 length plus payload per element), so treating them as a length table walks the cursor off the rails and produces astronomically large reads. Reading and skipping are accounted separately: key names and types are always read, string values over 256KB and string arrays are only walked past without copying, non-string arrays over 64 elements are skipped outright, and a cursor past the file size stops the read, because an out-of-range position means the cursor is wrong rather than that a limit should be raised. The displayed name goes Ollama tag, then general.basename plus size_label, then general.name, then the file name, with the full string kept for the hover text since it usually carries internal build numbers; mmproj files are now identified by general.type as well as by name. A short read (truncated file) no longer feeds half a value into the parser. On the reference machine, reading one file header parses 44 keys in about 16ms, and a whole scan takes about 31ms. From 0.4.7 the smoke test no longer fails the first time and passes the second. The old code gave both request attempts one shared AbortController and one shared 30-second timer, so the total budget covered both rounds: a cold load of a large model burned it in round one, the abort left the signal permanently aborted, and the second round (the Ollama-style reasoning_effort: none retry) could not even be sent - which is exactly why the first click returned This operation was aborted and the second succeeded. Now each attempt gets its own controller and timer, a timeout retries once automatically (retryOnTimeout, on by default) because a first-request timeout almost always means the weights are still being read from disk, and non-timeout errors (refused connection, HTTP error) are not retried at all. Timeouts report themselves: the message says the wait exceeded smokeTimeout, whether a retry already happened, and points at the new smokeTimeout setting (seconds, default 60, clamped to 5-600) - port responding is not the same as tokens coming out, so it is a separate knob from launchTimeout. Unrelated: an example naming one specific model was dropped from the panel copy, since not every machine has that model downloaded, and the explanation now describes the mechanism instead. From 0.4.8 adopting a plain .gguf no longer registers an unrelated Ollama model. The port field used to follow the last click rather than the file: selecting an Ollama entry moved it to 11434, selecting a plain .gguf did not move it back, so a 11434 left over from an earlier click stayed. The server accepted it as given, probed the port, found Ollama answering and borrowed it (the skipped-launch log line about the port already answering is this path), and then the model id - derived from the file name - could not be matched against Ollama's /v1/models, where the fallback was to grab the first entry in the list. Embedding models tend to sort first, so the adopt registered a model the user never selected. Now the port follows the file: a plain .gguf landing on another engine's port (ollama 11434, vLLM 8000, LM Studio 1234) is moved to 8090 and the panel says so. Only ports belonging to a known engine are overridden - a hand-typed 9999 aimed at your own llama-server is left alone. The model-id fallback no longer guesses either: it takes the only entry when there is exactly one (a single-model backend may legitimately report a different id than the file name) and otherwise reports an error listing the ids it did see, rather than registering a wrong provider. From 0.4.9 the panel can no longer vanish. It used to be possible for the whole section to disappear from the plugin detail page while the host stayed up, the HTTP bridge kept answering, and no crash log appeared. The cause was a scope error: a helper was declared inside one component and referenced from another, which is valid JavaScript and passes a syntax check, so it only threw at render time - when the user pressed the scan button and React ran the state updater. The host wraps every slot entry in an error boundary whose fallback is an empty anchor, so the throw was contained silently: the panel was not uninstalled or crashed, it was replaced by an empty div. The helpers now live in the module scope both components share. The client half also gained its own test - a minimal React stand-in loads the real panel module, really calls the components, really runs the state updater at render time, and presses the scan button, so this class of error can no longer reach a release.
dsh-unified-desktop
nalikip/dsh-unified-desktop
One MCP tool surface over two Windows desktop-automation backends (dsh-computer-use-win and windows-mcp). Both must be installed first: dsh-computer-use-win as a DSH plugin, windows-mcp via uv tool install windows-mcp --python 3.14, then run the bundled scripts/patch-backend.mjs once (it rewrites the local windows-mcp source; without it screenshot/read_webpage/launch_app fail) and disable the standalone wincu mount. It merges their 44 tools into 29 by dropping 10 duplicates, 4 dangerous ones (PowerShell, Registry, FileSystem, Process) and 1 tool-management entry, and blocks dangerous calls before forwarding through a reversibility-graded guard.
dsh-memory
bangbang-03/dsh-memory
Long-term memory scoped to the DSH workspace. A session opens with a one-line-per-memory index - bodies are fetched on demand by the `memory` tool, so the same character budget holds about three times as many entries - and when a turn ends a purely local search over the last user message pulls the matching memories into the next briefing at no model cost. Optional auto-capture distils a stretch of conversation into structured memories; in draft mode they wait in an approval queue instead of landing directly. A portrait layer keeps a workspace and a global picture, rewritten from the journal and the records, with the same draft-or-direct choice. Settings → Memory holds the archive, list and timeline views, a health check (duplicate clusters, stale entries, broken links), the supersede version chain with rollback, and a read-only scan that imports existing WorkBuddy / Qoder memory files.
dsh-connect-workbuddy
ethanweave/dsh-connect-workbuddy
Runs a local OpenAI-compatible gateway so DSH can use the model quota of a logged-in Tencent CodeBuddy (WorkBuddy) account: one-shot installer with SHA256 verification, logon scheduled task with duplicate-launch guard, manage.ps1 for status/start/stop/restart/logs/update/uninstall, and an agent skill that drives setup and troubleshooting.
dsh-screenshot-capture
wangzhanchao883/dsh-screenshot-capture
Point-and-shoot screenshot capture: a system floating window turns a new screenshot (or copied image) into an Obsidian note with comments and key-point marking, instant OCR through the host DSH multimodal model (no extra API key needed, Tongyi Qianwen as fallback), a duplicate-screen guard, per-day merging, and an optional evening AI organization pass.
dsh-context-guard
person-clouds/dsh-context-guard
Keep DSH tool output within a context budget: archive full output to disk, deduplicate repeated results, show only bounded diffs, ask before risky shell commands, and provide read-only inspection scripts.
alice-memory-llm
cycocyco/alice-memory-llm
Long-term memory for DSH where an LLM decides what to keep: turns pool while the session is idle, then the main chat model distills them into one-line entries in a single batched call; vector dedup runs before insert, so near-duplicates and saturated topics are dropped; recall uses hybrid vector plus FTS5 search over tiered hot, warm and cold storage; ships remember, recall, forget and prune tools. Zero dependencies, Node only.
dsh-plugin-lrc
zhibaiyingchuan/dsh-plugin-lrc
Connects DeepSeek Harness to LRC's local structured memory: distills conclusions every 8 turns, drops near-duplicates by character-level Jaccard similarity (≥ 0.85), and writes them to a local LRC store. Five tools — lrc_recall, lrc_forget, lrc_remember_now, lrc_handoff and lrc_status — cover on-demand recall, retracting wrong entries, immediate writes, session handoff and a status line. On first run it downloads a per-platform LRC binary (win32-x64, darwin-arm64 or linux-x64, SHA256-verified) and starts an isolated instance, so the LRC desktop app is not required; data stays in $DSH_HOME/cache/lrc-plugin/data unless LRC_BASE points at an existing instance.