@kai232/dsh-rag
Local knowledge bases with explicit, source-backed retrieval for DeepSeek Harness Web.
88 results
Local knowledge bases with explicit, source-backed retrieval for DeepSeek Harness Web.
Reproducible local experiment matrices for DSH profiles
KnowLP-RAG: dual knowledge-graph retrieval for Markdown notes — DeepSeek Harness (dsh) bundle: MCP server + native Cordis plugin
Benchmark-driven self-evolution plugin for DeepSeek Harness: evaluate, optimize, snapshot, accept or roll back.
Deterministic, CI-safe golden-output evaluation for DeepSeek Harness
Notebook Knowledge Studio — NotebookLM-like local knowledge workspace for DeepSeek Harness: Notebooks, Sources, source-grounded Chat with citations, and Studio artifacts (Mind Map / Reports / Flashcards / Quiz / Infographic / Slide Deck / Data Table), stored as OKF v0.2 markdown under the workspace.
Thin on-demand context-retrieval layer for DeepSeek Harness: context_query / context_slice / context_grep tools over the sessionQuery and subprocess seams.
AI-driven partial wave analysis for DeepSeek Harness: physics-gated config editing, ctpwa fit execution, numeric evaluation, and goal-driven iterative convergence. Physics knowledge (PDG-2026) as pure functions; DSH integration as a thin pwa_* tool plugin.
Replay real DeepSeek Harness turns against Standard, Minimal, Anchored, or plugin candidates with frozen request-surface evidence
Skill-trigger evaluation: an LLM judge recreates the DSH skill catalog and measures how reliably a skill description routes matching queries.
DeepSeek Harness bundle: a Chrome browser for agents (browser_goto / browser_evaluate / browser_screenshot and 12 more) plus a no-key net search tier (browser_search / browser_fetch / Reddit tools, ported from pi-lynx) with a live two-way pane inside the DSH Web GUI. The browser engine is a fork of zenbu-labs/terminal-browser with the React Ink terminal UI replaced by a DSH tool surface and web pane.
Cross-session long-term memory for DeepSeek Harness: dual-scope SQLite memory graph, hybrid retrieval, provenance audit, and a knowledge flywheel.
dsh bundle: local semantic index over workspace documents (MD/PDF/DOCX/TXT) with FTS5 BM25 + local embedding hybrid retrieval, hit citations with line numbers, and incremental updates.
Two-stage intent and evidence review, deterministic completion gates, and bounded repair for DeepSeek Harness agents
Persistent memory for DeepSeek Harness: per-conversation notes, selective injection, project memory, semantic Vault search (bge-m3, free optional key) and automatic retrieval — agents stop forgetting.
dsh 插件开发知识库:DeepSeek Harness 官方文档站完整 Markdown 镜像(中英双语 184 页)+ 58 篇仓库补充开发参考 + 主题导航与 245 文件全量搜索索引;以 dsh-plugin-dev-kb 技能让 agent 在插件开发任务中自动加载定位文档。纯数据插件,零运行时依赖。
OpenAI Responses API-backed web search provider for dsh: registers into ctx.web (web_search tool with server-side retrieval), plus a Web settings card for endpoint/model/key. Hot-pluggable — mounted via profile bundle patch, no dsh source changes.
Audit an agent harness against the harness-evaluation criteria, with machine-enforced evidence validation.
A local DeepSeek Harness north-star guard with explicit AI indicator evaluation and task alignment context.
Compare multiple coding models side by side in DeepSeek Harness with isolated Git worktrees
DSH agent preset for rigorous strategy live-deployment testing/evaluation. Retest.
Semantic long-term memory for DeepSeek Harness: embedding-based retrieval over a persistent cross-session memory store, with model-facing tools, proactive per-question recall, automatic conversation summarization, and auto-selected embedding provider (apiKey present → API, otherwise local).
Community visual workflow and multi-model evaluation plugin for DeepSeek Harness
Check which DeepSeek Harness plugins actually loaded, and look up public evaluation boards on trapstreet.run