Plugin
Sfoglia, filtra e installa i plugin di DeepSeek-Harness.
320 plugin trovati
dsh-tool-vision
gloryxpnv/dsh-tool-vision
Visione strutturata local-first per agenti solo testo: le immagini vanno a un VLM locale compatibile OpenAI e tornano come prove JSON (sintesi, OCR letterale, regioni di layout, entità/relazioni, colori, incertezza esplicita), con fallback anti-allucinazione e un ponte opzionale di incolla/carica; costo cloud zero, le immagini non lasciano mai la macchina.
dsh-memory-porter
shiye-10pages/dsh-memory-porter
Porta la tua cronologia Claude / ChatGPT esistente in DSH: importazione istantanea senza token del memories.json dentro un export di Claude, trascrizioni locali di Claude Code senza bisogno di esportare e conversazioni distillate dal modello già configurato in DSH. Ogni memoria porta prove letterali verificate rispetto alla fonte dal codice — una parafrasi viene scartata e conteggiata, e le prove trovate solo nei turni dell'assistente vengono sempre indirizzate alla revisione invece che archiviate.
dsh-failure-capsule
yiharvest/dsh-failure-capsule
Capsule di evidenza dei fallimenti local-first per le sessioni di DeepSeek Harness
project-flow-explainer (dsh-plugin)
hieupt1904/project-flow-explainer
Explains a codebase flow as a self-contained offline HTML diagram with file:line evidence, and gives the agent seven code-intelligence tools (index, search, context pack, impact, verify, refresh, flow) plus a /project-flow-explainer slash command. Static analysis only: it never runs the target repository and makes no network calls.
dsh-research-harness
oseas031/dsh-research-harness
Research workbench with evidence-bound findings, guard receipts, completion gates, a CLI, ten bundled skills, and paired skill-evolution evaluation.
deepseek-harness-plugins (evidence-ledger)
singapo/deepseek-harness-plugins
Verifies numerical claims against pinned local data or World Bank population data and blocks delivery when checks fail.
dsh-think-twice
sshooting/dsh-think-twice
Question-first DSH composer review that asks up to two evidence-backed questions without rewriting the draft.
dsh-zentao-buglens
zhan1024/dsh-zentao-buglens
BugLens turns annotated screenshots and page evidence into guided bug reports, suggests frontend/backend triage and recent assignees, and creates ZenTao bugs with attached materials.
dsh-TradingAgents (npm)
vicfei/dsh-tradingagents
Multi-agent investment committee for DSH - analyst, bull, bear, risk and PM debate where every claim carries an evidence tag, abstention (PASS) is a first-class verdict, Buffett/Munger/Lynch/Wood guest seats critique the debate itself, and zero-key market data (CBOE quote/K-line/options-IV + SEC 10-K + nasdaq news, Yahoo-free) feeds an auditable IC-MEMO - 6 to 10 model calls per session.
endo-decision-workbench
flying-ovaltine/endo-decision-workbench
Endocrinology decision workbench for frontline endocrinologists. A differential table ranked on two axes (likelihood x harm if missed) where every row must carry counter-evidence, one of six evidence grades on each recommendation, branching workup paths and a treatment plan; emergency red flags collapse everything into three lines. The clinical rules ship as the skill endocrinology-decision-team.
dsh-skill-eval
carlslin/dsh-skill-eval
Scores every skill 0-100 from real execution telemetry (load success, follow-up tool errors, user corrections, ignored loads and cost) with 95% confidence intervals, turns that evidence into fixes the agent can act on, and copies one skill's current issues from the board as a ready-to-paste Markdown repair task.
dsh-factmesh
jidekaixin2dian/dsh-factmesh
Reconciles PowerPoint report figures with Excel evidence and reuses confirmed links in DeepSeek Harness.
dsh-tokeneconomizer (prune)
sonettogu14/dsh-tokeneconomizer
Hides the tools that stayed idle across previous sessions, decided once when a session starts from recorded evidence; it ships disabled, and the compression, skill and image tools are never hidden.
dsh-gated-workflow
invisiblezero/dsh-gated-workflow
A four-state workflow with program-level gates for DSH agents: discussion blocks file writes, implementation blocks out-of-scope edits, and verification requires evidence before checkmarks. Its HTTP routes have no DSH access check, so a web page open in your browser could call them.
dsh-oo-spec
jackguogit/dsh-oo-spec
Object-oriented design work run through four gates and nine judgement cards: one skill routes requirement, domain, design and implementation tasks; five tools search and read the bundled spec, resolve original-text anchors, build the ≤5-question block with defaults, and check a deliverable for missing slots, use-case parts, relation multiplicities, change windows and design↔code class parity with line-number evidence; a composer control opens the spec in-app. The gate checks are text heuristics.
dsh-deepsight
deeparchi-ai/dsh-deepsight
Adds a deep-research agent preset with the DeepSight methodology skill bundled, plus three tools that set an evidence contract, audit claim tiers along the provenance chain of every action item, and summarize the persisted ledger. Delivery is blocked while an action basis derives from unverified estimates or an unresolved conflict.
dsh-frontend-replay
zhan1024/dsh-frontend-replay
Frontend Replay records and replays frontend bugs, verifies fixes against explicit acceptance conditions, and exports browser evidence and Playwright tests.
dsh-opticforge
wbmgggg/dsh-opticforge
KB-first optical-engineering task assistant. opticforge_assist aligns a concrete problem with the 40-section / 324-entry knowledge base before any web search. 3.1.1 activates up to three topic playbooks (16 cards, 98 dimension declarations), retrieves evidence per required dimension, and returns a dimension-by-source coverage matrix, a gated plan with dimension IDs and cross-check steps, formula cards, a five-section report outline and acceptance checks. Answers explain principle -> formula/criterion -> approach -> steps, and never expose KB layer or entry IDs; sources are named as documents, standards, manuals or URLs. web_gate allows external search only for declared KB gaps, standard versions, vendor parameters or price timeliness. opticforge_verify checks KB alignment, present-only citations, dimension evidence, cross-relations, source tiering, KB conflicts, 30-day freshness and five-section presentation, and reports an adoption gate. Includes 308 concept cards, 6 generic dimensions, 19 formula cards and the opticforge-task DSH Skill. 308 entries are citable; 16 declared gaps remain, limited to A03, C01 and E01 verbatim sources. Five tools, zero runtime dependencies, no calculator, no network access and no action execution by the plugin. npm publication is on hold; install from GitHub with node scripts/install-local.mjs.
dsh-plugin-models-fingerprint
darkhe55/dsh-plugin-models-fingerprint
Monitors DSH model routes by comparing protocol fingerprints and behavior statistics across sessions, with per-turn cards, route equivalence, official baselines, and drift/quality signals. Fingerprints are evidence, not proof of model identity. Its HTTP routes have no DSH access check, so a web page open in your browser could call them.
dsh-task-readiness
kelvin0304ykx-coder/dsh-task-readiness
Checks whether the current session can actually run a task before starting it: tool admission for the current step, model route, workspace write boundary, document engine and MCP visibility, reported per capability with evidence, plus an opt-in minimal probe and a recheck after an approved fix.
daily-report-skill
wonderlawrence/daily-report-skill
Writes Chinese daily and weekly reports from the GitHub pull requests you took part in, keeping only your own commits inside the date range and adding Claude Code, Codex and DSH session evidence, with a JSON evidence snapshot saved beside the report so each claim can be traced back to a pull request, commit or session.
dsh-evidence-audit
sas-159/dsh-evidence-audit
Checks the agent’s own claims against the tool output saved in the same session: each sentence carrying a path, a number with a unit, inline code or a URL is graded supported, partial or no evidence, with the anchors it could not find listed. Coverage shows on the composer pill. Heuristic and local — no LLM call, no receipt log, nothing uploaded.
dsh-rtl-flow
wanghao815/dsh-rtl-flow
Guides RTL design and verification in DeepSeek Harness, with project scaffolding, evidence checks, and OpenROAD/ORFS setup.
dsh-project-evolution
triagmas/dsh-project-evolution
Captures verified project work as masked evidence and reviewable skill proposals; accepted skills become native project skills with version history, diff, rollback, and opt-in global publishing.