플러그인

LLM-Wiki형 세션 간 Markdown 메모리와 SOUL.md 페르소나. 기본 주입은 첫 사용자 메시지 뒤 활성 세션에만 적용. remember/recall/consolidate/forget. session format v4(공식 DeepSeek Harness 0.1.7+) 및 source kind plugin:memory 필요. v0.13 engines.dsh는 공식 0.2.1-rc.5 계열만 허용하고 구버전은 거부; Review now의 POST /api/memory/control action growth-now는 floor 준수 recap 1회 전송. v0.14 공용 세션 gate(chip, /memory on|off|status, 설정): off는 주입/digest/회상/last_access 중지, 삭제 없음. disabledWorkspaces로 프로젝트 전체를 기본 off로 설정하고 세션 선택은 7일간 유지. v0.15 Basic soul은 SOUL.md 관리 블록만 갱신하고 나머지 보존, BOOTSTRAP 완료, 기존 soul 반영 및 없을 때 생성 알림. checkup 보고서(우선순위 작업 목록).

Zero-residence 컨텍스트 엔진: 대용량 도구 페이로드를 라이브 컨텍스트에서 축출하여 쿼리 가능한 저장소로 옮기고 결정론적 포인터만 남기므로, 턴당 프롬프트 토큰이 대화 기록과 함께 계속 늘어나지 않습니다. 마스킹된 내용은 도구 호출 ID 또는 세션 seq로 필요할 때 세션 로그에서 그대로 복원됩니다.

zero-knowledge 릴레이를 통해 휴대폰에서 도구 호출을 승인하고, 스트리밍 출력을 확인하며 질문에 답할 수 있습니다. 페이로드 수준 종단 간 암호화를 사용합니다.

Claude Pro/Max 멤버십으로 DeepSeek Harness에 로그인하고(API 키 불필요), Anthropic 요청 이미지를 긴 변 2000px로 축소하고, /compact 요약기 페이로드를 슬림화하고, 5시간/주간 플랜 사용량과 DSH 메시지 수를 표시합니다.

Windows 전용입니다. DSH를 시작할 때마다 별도 설정 없이 원신 설치 여부를 확인합니다. 설치되어 있으면 게임을 실행하고, 업데이트 대기 등으로 바로 종료되면 miHoYo 런처로 전환합니다. 미설치 상태면 런처 설치 파일을 바탕 화면에 내려받되, 예상 용량 약 224 MiB에서 15 MiB 넘게 벗어난 링크는 거부합니다. 다운로드에 실패하면 공식 다운로드 페이지를 엽니다.

Claude Code 스타일 생명주기 훅: 모델 도구 호출 전후에 stdin으로 JSON 페이로드를 전달받아 설정된 셸 명령을 실행; pre-tool 명령이 0이 아닌 값으로 종료되면 호출을 차단

Adds a { } button under each assistant reply in the DSH Web GUI that opens the exact request sent to the model for that reply: model, call parameters, messages, tools, and provider-reported token usage.

Policy-layer guard that checks tool calls before they run: a read whitelist, a sensitive-path deny list (credentials, browser profiles, chat history, other users' home directories), dangerous-command blocking, subagent isolation, and a static scan that quarantines executables. Every denial is appended to an audit log. It is a text-and-path filter, not a sandbox or antivirus: dynamically built paths, encoded payloads and packed binaries can get past it, and it does not replace the OS antivirus.

Precomputes the compaction summary in the background and commits the held result when pressure compaction fires, removing the foreground model round-trip. Registers itself as the session compaction engine, pins the span it summarised so the payload fingerprint still matches at commit time, prunes tool results before committing, and falls back to the stock foreground path on payload drift or context overflow.

외부 데이터베이스와 문서 코퍼스를 모델 호출 가능 도구로 등록합니다. 소스별 접근 경계, 값을 저장하지 않는 자격 증명 참조 원장, payload 대신 fingerprint와 콘텐츠 해시만 보관하는 fetch 원장을 제공합니다.

DSH Host 재시작으로 중단된 턴을 재개합니다. 세션이 활성화된 뒤 agent activation을 hook하고 복구된 turn/end를 읽습니다. 부작용의 성공/실패를 가정하지 않고 실제 상태를 먼저 확인하는 continuation을 보냅니다. AUTH, QUOTA, context overflow 같은 영구 오류는 건너뛰며 재시작 간 예산, 적응형 backoff, 최소 context payload를 사용합니다.

지원되는 파일 변경에 대한 턴 단위 설정 가능 예산: 도구 실행 전에 대상 파일 수, 변경 호출 수, UTF-8 페이로드 바이트를 제한합니다.

모든 업스트림 모델 API 페이로드를 JSON으로 캡처(디버깅 및 관측성)

Approval frame for DSH tool calls: a rule table with four outcomes (deny, ask, AI review, allow) across tool, command, file and network scope, then AI fast/deep review and human approval, a hash-chained audit trail and a settings UI. Target files are inspected by metadata only (no file contents are opened); the reviewer sees the command, targets and clipped write payloads, with secrets redacted before anything leaves the machine.

Look up Chinese companies inside the conversation: name a company and get its business registration, shareholders, executives, outbound investments and registration changes, plus judicial and operating-risk records such as enforcement, dishonest judgment debtor, consumption restriction, abnormal operation, serious illegality and administrative penalties (11 dimension groups). Results render as conversation cards with fields and the raw payload on expand, and the API key is saved from the plugin page into the Harness credential store.

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.

이벤트 기반 trigger 플러그인(dsh-proactive 알람시계 옆의 초인종)입니다. 규칙별 debounce 창과 glob filter가 있는 재귀 fs watcher, 규칙별 token이 있는 loopback 전용 webhook endpoint, host agents service로 대상 session을 깨우는 fire rule을 제공합니다. 활성 session은 직접 후속 메시지를 받고, 비활성 session은 먼저 재개합니다. paths/count/payload 템플릿 변수를 렌더링한 prompt를 주입합니다. DSH 재시작 없이 규칙 파일을 hot-reload하고 모든 trigger를 감사 가능한 JSONL 기록에 추가하며 trigger_list/trigger_history로 조회합니다. loopback health route에서 설치 및 실행 상태를 확인합니다. 다운로드 폴더 자동 정리, 로컬 스크립트가 보낸 CI/모니터링 경보 진단, 로그 오류 감시에 활용할 수 있습니다. Pure JS, 빌드 단계 없음, MIT.

DSH Web UI 세션 헤더 또는 rea_mcp_toggle agent tool에서 rea-mcp server를 켜고 끕니다. REA가 꺼져 있으면 170개 mcp__rea__* tool schema를 매 요청에 보내지 않아 tools payload가 약 722 KB에서 약 95 KB로 줄고 REA 하위 프로세스도 종료됩니다. 선택은 프로필 범위 state file에 저장되며 시작 시 row-level disabled expression으로 복원됩니다.

일반 자연어 요청을 실시간 워크벤치 창으로 변환한다. 사이드바 패널이 몇 가지 확인 질문을 하고 모델은 단일 파일 HTML 결과물을 작성해 크기 조절 가능한 플로팅 창이나 독립 브라우저 탭에 표시한다. 선언된 창 동작은 페이로드 없이 정보 수준 프로젝트 저널에 기록한다. 워크벤치를 가져오고 내보내거나 마켓 인덱스에서 설치할 수 있다.

프록시 provider로서 전송 payload를 범주별 op 체인으로 재구성한 뒤 upstream에 전달합니다. 턴 사이 prompt prefix를 단조롭게 유지해 변경되지 않은 기록에 전체 비용이 다시 청구되지 않게 합니다. op 이름은 exec/http/inline transport를 뜻하며 사용자 정의 알고리즘을 받습니다.

모터 설계 AI 보조 도구입니다. 요구 분석→매개변수 식별→후보 행렬→상위 10개 검증 순으로 진행합니다. L0는 밀리초 수준의 광범위 필터(순수 JS, solver 의존성 없음)로 후보 순위를 매기고 L1/L2 정밀 도구에 넘길 데이터를 생성합니다. 로컬 에이전트 프리셋과 Cordis 플러그인으로 사용할 수 있습니다. 이번 버전에서 L1(RMxprt)과 L2(Motor-CAD)는 예약 상태이며 l0 초과 단계는 즉시 거부됩니다.

DeepSeek Harness용 에이전트 안전 컨텍스트 압축기입니다. 오류, 파일 경로, JSON 데이터, 명령 결과 같은 중요한 도구 출력을 의미적으로 보존하며 기계적으로 잘라내지 않습니다.

DeepSeek Harness 도구 결과를 위한 사전 콘텐츠 정책입니다. 정규식 규칙에 맞는 텍스트를 바꾸거나 필드 경로를 기준으로 예측하기 어려운 데이터를 제거·잘라내 정규화·표시·저장 전에 처리합니다. dsh-plugin-redact의 upstream 대응 플러그인입니다.

도구 결과용 선제적 콘텐츠 정책입니다. 결과가 정규화·렌더링·저장되기 전에 일치한 텍스트를 다시 쓰고 경로에 따라 필드 전체를 버리거나 자릅니다. 검토되지 않은 페이로드가 처리 단계에 들어가지 않게 합니다. 기본은 비활성 상태이며 내장 범주/도메인 차단 목록이 없습니다.