Plugin
Sfoglia, filtra e installa i plugin di DeepSeek-Harness.
95 plugin trovati
dsh-canvas-plugin
answertsundere/dsh-canvas-plugin
Canvas dei rami di conversazione: grafo genealogico delle sessioni a tutta area, diramazione da qualsiasi messaggio, esecuzioni parallele in streaming, confronto deterministico dei rami e vista web dello sciame di agenti per sessione.
dsh-plugin-background-compaction
jisiyong-a/dsh-plugin-background-compaction
Precalcola in background il riepilogo della compattazione e conferma il risultato trattenuto quando scatta quella per pressione, eliminando un passaggio del modello in primo piano. Si registra come motore di compattazione della sessione, blocca l’intervallo riepilogato affinché l’impronta del payload coincida al momento della conferma, elimina prima i risultati degli strumenti e torna al flusso standard in primo piano se il payload cambia o il contesto trabocca.
dsh-sheet-diff
po-et/dsh-sheet-diff
Confronto CSV/XLSX in sola lettura tramite chiavi business, con decisioni di revisione offline, commenti e passaggi di feedback verificati alla fonte.
dsh-input-optimizer
xkjxziochu/dsh-input-optimizer
Riscrive la bozza del composer DSH in un prompt più chiaro ed eseguibile tramite le route LLM del harness, mostrando il confronto originale/ottimizzato prima dell’adozione.
dsh-plugins
youssefsiam38/dsh-plugins
Confronto affiancato nella UI web DeepSeek Harness: invia un prompt a 2–4 modelli in sessioni sorelle derivate da quella corrente e mostra risposte in colonne (schede sui telefoni) con latenza, time to first token, token e costo.
dsh-change-review-live
sujingkpo/dsh-change-review-live
Revisione modifiche live per DeepSeek Harness: traccia chiamate write/edit per sessione, elenca i file modificati sopra il composer e apre diff per riga in una scheda registrata nella barra laterale destra. Confronto con log sessione, versione iniziale o Git; ripristino per modifica/file, aggregazione subagenti e aggiornamenti SSE. Fork di cirelir/dsh-change-review ricostruito intorno alla sidebar destra.
dsh-project-memory
layzr114/dsh-project-memory
Polaris Memory: plugin di memoria dei progetti local-first per DeepSeek Harness. Usa `.dsh-memory.json` per progetto e offre gli strumenti `memory_read/recall/write/update/delete`, il comando `/memory` e un’interfaccia Settings. Include iniezione a livelli (profilo utente residente e ricordi P1/P2/P3 attivati dal contesto), richiamo semantico (IDF + sinonimi) e protezione rigida dei ricordi essenziali: rileva semantica di rifiuto, richiede autorizzazione e blocca l’eliminazione da parte dell’AI; l’utente può superarla dopo conferma.
wangtie-os
wangtie-010101/wangtie-os
Framework a schermo intero per app aziendali in DeepSeek Harness Web. La dashboard con barra laterale scura integra test di connessione DB e query SQL; knowledge base a due livelli (fatture e contabilità) con ricerca locale e import Word; confronto dei metadati ambientali e generazione ALTER SQL; dizionario dati; demo di report di ricerca nei log degli agenti; esecuzione di test case, note stile Evernote, utilità per sviluppatori e minigiochi.
dsh-product-subagent-console
jokasa7/dsh-product-subagent-console
Workbench multi-agente per conversazione DSH con pianificazione modificabile, osservazione delle sessioni figlie reali, confronto tra piano ed esecuzione e anteprime di ripristino basate su evidenze. La descrizione dichiara test con DSH 0.1.1-rc.2.
dsh-prompt-optimizer
lhf198/dsh-prompt-optimizer
Migliora la bozza chat corrente prima dell’invio, mostrando un confronto da confermare prima di applicare.
dsh-healthcheck
taxueseek/dsh-healthcheck
Controllo di salute dell'ambiente DSH in sola lettura: rilevamento di disco, memoria, latenza, gonfiamento di ~/.dsh e plugin obsoleti con report graduati e confronto delle tendenze rispetto alla baseline. Non elimina nulla.
dsh-fork-diff
chouyong/dsh-fork-diff
Confronto affiancato in sola lettura di cronologie di conversazione DeepSeek Harness correlate, inclusi messaggi, chiamate agli strumenti, risposte, tempi e utilizzo.
paper-dual-reducer
mount9527/paper-dual-reducer
Paper dual-reduction workbench for Chinese Word documents: reads a .doc/.docx in place without changing it, scores it with a local Chinese AIGC detector, runs a nine-dimension AI-trace diagnosis, and turns the result into batch rewrite prompts for any agent; the pasted-back edits are validated and applied through the paper-dual-reduce engine with tables, formulas, footnotes, TOC fields and citation marks left intact, then the frozen blocks are re-scored and the whole draft rolls back if its AI share rises. Adds a sidebar panel with a rewrite version chain of up to five passes, a three-column before/after comparison, downloadable per-version reports, and five paper_dual_* tools. The detector weights download on first run and are never bundled. Its HTTP routes have no DSH access check, so a web page open in your browser could call them.
dsh-github-repo-search
zhaodezhang9998/dsh-github-repo-search
Read-only GitHub search for agents: repositories with a 1000-result paging window, code contents, issues and pull requests, and a derived trending query ranked by recent pushes and stars. Results collect in a right-sidebar panel split into repositories, code and issues, where cards expand to language mix, latest release and README, and can be starred across sessions, kept as a cross-session query history, or put two or three side by side for comparison. A token from the DSH credential store is sent only to github.com, api.github.com and hosts listed in trustedHosts.
dsh-token-tally
bioaievolu/dsh-token-tally
Token & cost dashboard: model-by-time stacked bars (absolute / 100% share), per-model trend lines, donut charts, rankings, a model comparison table (tok/s speed, price per M tokens, cache hit rate, P50/P95 latency) and a 53-week heatmap — aggregated locally, zero third-party runtime deps. Its HTTP routes have no DSH access check, so a web page open in your browser could call them.
dsh-think-flow
cv588888888888888888888ju/dsh-think-flow
Chinese add-on fork of the dsh-think-flow sidebar chain-of-thought panel: while upstream folds each session's live reasoning into turn → step groups, this fork sends each step to a separate, configurable model (titleProvider/titleModel) and shows an 8–18 character Chinese title plus a 20–60 character Chinese explanation under every step, and a one-line Chinese title per turn. The main model is untouched — no system-prompt injection, no extra tools, no context changes; the plug-in just makes its own request. autoTitlesOn arms title generation as soon as a panel connects, and batches over 10 steps are split so long turns (100+ steps) no longer come back empty; steps that only call a tool (no reasoning) are summarised from the call itself, and a failed segment is retried up to three times with backoff.
dsh-loan-cal
aaalittlecloud/dsh-loan-cal
Strict-convention loan math for DeepSeek Harness: real APR from installment fees, amortization schedules, bank-statement audits, prepayment break-even, offer comparison and LPR repricing. Zero dependencies.
dsh-r-workbench
hougelangley/dsh-r-workbench
R data workbench with six tools: ingest CSV/TSV/text, Excel and xlsx-compatible WPS .et, ODS, JSON/NDJSON and ZIP files; clean and infer types; score whether the data can support analysis (PROCEED, CONDITIONAL or STOP); run group comparisons, correlations, association tests and calendar-month trends with effect sizes and BH correction; and write self-contained HTML, Markdown, Quarto source or DOCX reports. STOP blocks analysis and reporting.
dsh-weight-journal
begin2023/dsh-weight-journal
A daily weight journal for the Web GUI: one reading per calendar day, a monotone-cubic trend chart over 7/30/90-day and all-time windows with the entries' 7-reading mean and a goal line, weekly-mean comparison bars, a sidebar panel, a weight_journal tool (record, list, delete) that writes through the same store as the page, and a frame-wide reminder capsule that rotates through eight languages.
dsh-vellum
avenyq/dsh-vellum
Native cards inside DSH replies: when the model writes a vellum code block, it renders as a theme-following metrics, chart, comparison, checklist, steps, timeline, form, quiz or sortable table card; clicks, submits and answers come back to the conversation as the next message.
qianDao-photo-finder-skill
bananaive/qiandao-photo-finder-skill
Find photo candidates from the 2026 Qiandao Lake cycling race by bib number and visible gear, with local helper scripts and album snapshot comparison. Needs the event's photo album link; snapshot comparison uses Python 3.10+, bib OCR is macOS only (Swift and Apple Vision), and the optional gear ranking needs Pillow, PyTorch, Transformers and a local DINOv2 model directory.
dsh-featured-market
ghostmss86-afk/dsh-featured-market
Curated DeepSeek Harness plugin market with a daily heat leaderboard, bilingual descriptions, 2-3 plugin comparison, compatibility checks and README summaries.
dsh-xhs-style
refreshing-czq/dsh-xhs-style
Adds a requirements-discovery gate before product selection, version choices and competitive analysis: 2–3 clarification rounds with 2–4 decision-changing questions, validation of real pain points and hard constraints, comparison of original, alternative and minimum-cost routes, and a confirmation card before research and copywriting. Confirmed reports keep the Xiaohongshu-style structure, concrete steps, audited images and source links. Ordinary lookups and explicit user skips bypass the questionnaire. Host-only static prompt sections with no config, client UI or dependencies.
dsh-prompt-optimizer
nyco1d/dsh-prompt-optimizer
One-click prompt optimizer next to the composer, with a side-by-side diff panel and per-mode settings. Output length follows your request: ask for a short version and it writes a 350–650 character brief, otherwise it expands to as many as 1400 characters. The prompt templates are original; the official-template comparison mode needs a local Prompt Optimizer installation.