18 results
独立 dsh 插件：会话内初始化/接入本地王国，Territory/Binding CRUD + Task 治理闭环（Worker Claim ≠ Fact）+ Execution 生命周期 + v0.8 Persistent Governed Worker（长期 Session / Territory Affinity / Capability Gate / Execution Lease / Dispatch Evidence / Reconciliation）+ 本地 GUI 通道
DSH Research OS — a fully automated scientific research plugin for DSH (DeepSeek Harness): survey, idea, experiment contract, durable runner jobs, claim-evidence ledger, manuscript and release bundle.
Embedded C/C++ firmware development toolbox — 4 agents, 8 skills covering FreeRTOS, ISR, NVM storage, Keil MDK, ARMCLANG, HardFault, state machines, architecture, LVGL patterns, and claim fact-checking. Ships a native DeepSeek Harness (dsh) bundle that injects the session-start gate into the first model step.
Out-of-tree dsh provider plugin: a DeepSeek gateway route that claims image input and transparently describes pasted images through a configured vision-language model (e.g. Qwen-VL) before the text-only DeepSeek wire sees them.
Cross-session agent-to-agent messaging for DeepSeek Harness: claims, verification and a decision ledger, so two agent sessions don't repeat, contradict or deadlock each other.
Write in parallel, never overwrite — file claim / protection for concurrent DeepSeek Harness (DSH) sessions working the same workspace: claim/release, heartbeat stale takeover, and an async pending merge area (git 3-way merge)
Popper: a falsification-driven correction loop for agent sessions — claim commitment, gate falsification, and mutually exclusive hypothesis revision, with an append-only evidence ledger
Verification toolkit for DeepSeek Harness agents: evidence-based claim checking against workspace files with line citations, config file validation (JSON/YAML), HTTP URL status checks, npm package checks and GitHub repo submission-readiness checks — complementary to dsh-plugin-gate
Kestrel — adversarial deep-research engine for DeepSeek Harness. Mechanically-verified citations, independent-corroboration scoring that exposes syndication laundering, prosecutor/defender claim tribunal, and a compounding evidence graph.
理赔流程
Search arXiv for ML/DL/RL papers and check claims against real abstracts.
Routine taskboard for DSH: a dedicated registry of long-lived scheduled infrastructure placards (windows tasks / cron / ai sessions) with dual-source health lamps and artifact freshness, plus a lightweight claimable work-order board.
Tiered task queue for DSH↔Hermes coordination: queue.json is the single source of truth; tasks are enqueued by DSH mechanisms (review/lesson/reset) and consumed by the Hermes-side cron with a lease/claim model. Also provides the busy-mutex for serializing Hermes calls from the approve fast path.
怀疑驱动开发：每个非平凡决策在落地前接受新鲜上下文的对抗性审查（CLAIM→EXTRACT→DOUBT→RECONCILE→STOP）。受 addyosmani/agent-skills（88k★ MIT）启发。
多模型候选评审管线:调度智谱 GLM / Claude(CLI 无头,可走中转)/ Gemini(OpenAI 兼容口)/ Codex 四视角评审成长 artifact(教训/规则/审计板),输出严格 JSON(verdict/strengths/risks/suggestions/cannot_claim),结果只作 candidate_review_input,永不替代本地证据。适用:用户要求外部交叉评审、周期性仪式前自动评审、怀疑某条结论过度推广时。
AI deep-reading assistant for DeepSeek Harness and dsh-TUI: extract traceable claims, evidence, confidence levels, knowledge maps, and review questions from articles, books, PDFs, files, URLs, or pasted text.
车险理赔决策
DeepSeek Harness bundle: the scopus-scholar agent preset — Scopus-powered academic literature search and analysis with a claim-verify-plan research loop