FS-shaped internal URL schemes (conflict://, pr://, issue://, …) resolved through one registry behind the read/grep/write tools — a standalone service plugin for DeepSeek Harness with no upstream harness changes required.
Agent 협업
$ npx -p @deepseek-ai/dsh dsh plugin --profile web add github:hy-sde/dsh-internal-urls#0988f08b803409c6916928c321ce3e9bd5d59847&path:packages/internal-urls
Host-wide model-slot admission control (ctx.modelSlots) decided FIFO at the llm/stream waterfall: a shared budget over every model call so 2-3 concurrent local inference slots stay predictable, as a standalone npm package.
Host-plane wiki-graph service over the Logseq CLI: one typed seam (ctx.wikiGraph) the web UI and host API read and write the LLM-wiki graph through, headlessly and without the desktop app.
Agent-curated long-horizon memory (ctx.memory) for DeepSeek Harness: durable project-scoped memory banks with a provider registry and a shipped local backend — a standalone plugin with no upstream harness changes required.
메모리 및 지식 · Agent 협업
$ npx -p @deepseek-ai/dsh dsh plugin --profile web add github:hy-sde/dsh-memory#133522aaf655c0b0475b46736e3e2599efd694f6&path:packages/memory
The OpenWiki 0.4.3 deterministic engine (MIT) as an in-process library for DeepSeek Harness: resumable page-job lifecycle with durable .run.json/.page-manifest.json/.last-update.json, Grounded Claims store/session/evidence resolver with .claims/ sidecars, OKF v0.2 front matter validation/repair + index sync + generated provenance, Mermaid + wiki-link validation, behind a minimal WikiFs filesystem seam — no external openwiki CLI and no DeepAgents coupling.
The durable-agent engine for DeepSeek Harness: a host-plane cordis service mounting @earendil-works/pi-durable (conversations, exactly-once submissions, fork-at-entry, tasks) over a SQLite storage backend with an OpenAI-compatible relay provider.
Agent 협업
$ npx -p @deepseek-ai/dsh dsh plugin --profile web add github:hy-sde/dsh-pi-durable#978844c5660040e7ad6f2830e9b1cd527abe3e93&path:packages/pi-durable
The `session_health` tool for DeepSeek Harness — per-session health intelligence (outcome classification, tool-health signals, prompt/workflow heuristics, context pressure, A–F health grade) computed in-process from the session store, plus a standalone CLI over the same engine — a standalone plugin with no upstream harness changes required.
Agent 협업 · 생산성 도구
$ npx -p @deepseek-ai/dsh dsh plugin --profile web add github:hy-sde/dsh-session-intelligence#235fed024e4c57a85120028d63a6a8bf990d9da0&path:packages/session-intelligence
session:// internal-URL scheme handler: read a session transcript, read one event as JSON, list sessions, and search past history through internal-URL-aware read/grep tools — standalone for DeepSeek Harness.
Agent 협업 · 생산성 도구
$ npx -p @deepseek-ai/dsh dsh plugin --profile web add github:hy-sde/dsh-session-url#27316a2000e8a5408c6bf0835df04ec7b927daa8&path:packages/session-url
The ast_grep (structural code search) and ast_edit (structural rewrite) tools for DeepSeek Harness over the packaged ast-grep native engine — a standalone plugin with no upstream harness changes required.
Agent 협업 · 개발 도구
$ npx -p @deepseek-ai/dsh dsh plugin --profile web add github:hy-sde/dsh-tool-ast#7a398025e5fac6a6fa97acf637857ed7c8773a3a&path:packages/tool-ast
Model-facing Automic Vault tools (av_scan audit, av_doctor hardening check, av_catalog, av_list) over the host ctx.av service — read-only, never returns secret values — plus an av:tools system-prompt section for DeepSeek Harness, a standalone plugin with no upstream harness changes required.
Model-facing browser tool (open/close/run over launch, CDP-attach and relay backends, ARIA-snapshot + screenshot observations) for DeepSeek Harness — a standalone plugin with no upstream harness changes required.
Model-facing codebase-memory CLI tools (codebase_list_projects/index_repository/index_status/search_graph/query_graph/trace_path/get_code_snippet/get_graph_schema/get_architecture/search_code/detect_changes/manage_adr/ingest_traces/delete_project) that run one-shot queries against the codebase-memory daemon from the terminal — the local alternative to the stdio MCP client row
메모리 및 지식 · Agent 협업 · 개발 도구 · 서비스 연동
$ npx -p @deepseek-ai/dsh dsh plugin --profile web add github:hy-sde/dsh-tool-codebase-memory#41dcb8943cc3de149262a687c6866291bae99be7&path:packages/tool-codebase-memory
The `debug` tool for DeepSeek Harness — a model-facing Debug Adapter Protocol tool with 28 operations (launch/attach, breakpoints, continue/pause/step, threads/stackTrace/scopes/variables/evaluate, disassembly, memory, modules, output, terminate) over the standalone DAP seam — a standalone plugin with no upstream harness changes required.
메모리 및 지식 · Agent 협업 · 개발 도구
$ npx -p @deepseek-ai/dsh dsh plugin --profile web add github:hy-sde/dsh-tool-debug#fc2e20715abb38582f56b65fa557cc4c58baffd3&path:packages/tool-debug
The rich `edit` tool for DeepSeek Harness (replace / patch / apply_patch / hashline modes) with embedded format-on-write and diagnostics — a standalone plugin with no upstream harness changes required.
Agent 협업 · 개발 도구
$ npx -p @deepseek-ai/dsh dsh plugin --profile web add github:hy-sde/dsh-tool-edit#7cdfcde520474abe24cc0a24af4f5272bb8fbe85&path:packages/tool-edit
Agent-scope shadow of the DeepSeek Harness read/write/edit fs tools that also resolves conflict://, pr://, issue:// URLs through ctx.internalUrls — the routing surface that makes internal URLs work on stock harness releases.
Agent-scope shadow of the DeepSeek Harness glob/grep discovery tools that also searches conflict://, pr://, issue:// resources through ctx.internalUrls — the search half of the internal-URL routing surface for stock harness releases.
Model-facing agentic git tools (commit analyze + commit_apply split execution with validation and a P2 review gate on push, review with P0-P3 verdicts, and worktree isolated per-task checkouts with durable leases) over the host ctx.git service — a standalone plugin for DeepSeek Harness with no upstream harness changes required.
Agent Graph supervisor tools: view_agent_graph, update_agent_graph, yield_agent_graph over a host-provided graph controller, with the orchestration:graph prompt section.
Model-facing library_search tool: free cross-ecosystem (npm/crates.io/Maven/Go/PyPI/RubyGems + GitHub) 'has this already been built?' search — exact-name ranked, stars/downloads as fitness signals, no API key required
Agent 협업
$ npx -p @deepseek-ai/dsh dsh plugin --profile web add github:hy-sde/dsh-tool-library-search#49184db16b5b81856420a19d61c78bb59ec4f6bd&path:packages/tool-library-search
Model-facing Logseq CLI tools (logseq_list/upsert/remove/query/show/search/graph/server) that drive a Logseq database graph from the terminal — headless-capable alternative to the desktop MCP bridge
Model-facing long-horizon memory tools (retain, recall, reflect, memory_edit, learn) over the ctx.memory service, plus first-turn project-memory prompt injection — a standalone plugin for DeepSeek Harness with no upstream harness changes required.
Model-facing OpenWiki lifecycle tools (openwiki_begin/openwiki_submit_plan/openwiki_next_page/openwiki_submit_page/openwiki_finish) over the in-process deterministic engine @hy-sde-org/dsh-openwiki, plus an openwiki:tools system-prompt section that wires codebase-memory tools as structural discovery — a standalone agent-plane plugin for DeepSeek Harness, no external openwiki CLI.
Model-facing durable-agent tools for DeepSeek Harness (durable_agent_submit / write / status / history / fork / abort) over the @hy-sde-org/dsh-pi-durable engine service.
Agent 협업
$ npx -p @deepseek-ai/dsh dsh plugin --profile web add github:hy-sde/dsh-pi-durable#978844c5660040e7ad6f2830e9b1cd527abe3e93&path:packages/tool-pi-durable
검색부터 설치까지
작업으로 검색
기능 키워드나 카테고리로 범위를 좁히고 최근 업데이트 또는 GitHub Star로 정렬하세요.
기능 살펴보기
소개, AI 분석, 원본 README를 읽고 플러그인의 기능을 확인하세요.
확인 후 설치
호환성과 의존성 범위, 설치 안내를 확인한 뒤 제공된 설치 명령을 복사하세요.
DSH Hub 자주 묻는 질문
DSH Hub란 무엇인가요?
DeepSeek Harness 커뮤니티 플러그인 디렉터리입니다. 소개, 소스 링크, 설치 안내를 모아 확장 기능 검색과 비교를 돕습니다.
등록된 플러그인은 호환성 검증이 끝났나요?
아니요. 상세 페이지는 기록된 호환성과 의존성 범위를 표시합니다. 알 수 없음은 호환을 의미하지 않으므로 설치 전 프로젝트 문서를 확인하세요.
어떻게 설치하나요?
상세 페이지의 설치 영역을 확인하세요. 명령이 있으면 Harness 환경에 복사해 실행하고, 없으면 프로젝트 README를 따르세요.