Plugin

Analizza le immagini della conversazione tramite prompt specifici per modalità (caption, UI, documento, grounding, topologia, ecc.) e inietta evidenze strutturate con primitive di coordinate (box, punti, riferimenti) come testo, con caching a livello di sessione per il riutilizzo durante replay e compaction.

Consente a un agente DeepSeek Harness locale di utilizzare BailingHub per interrogare dati, aggiornare record ed eseguire azioni consentite in un negozio online, SaaS o altro sistema amministrativo.

Albero dei file nella barra laterale per DeepSeek Harness: un pannello file con un albero lazy con icone per estensione, azioni di copia percorso e invio all'agente.

Animale domestico balena cyber per DeepSeek Harness: una balena fluttuante, alimentabile e personalizzabile nei colori sopra shell.overlay che riporta uso di token, quota, interazioni e conteggi di sessione — un unico bundle dsh-plugin che porta il cervello di chat dell'host e la superficie del browser.

Icone in stile Fluent automatiche e personalizzabili dall'utente per DSH Settings e azioni della sidebar, con pacchetti SVG offline e conservazione delle icone originali.

Centro competenze nella GUI web di DSH: sfoglia le competenze caricate raggruppate per origine (system / project / user / custom / runtime), con creazione, attivazione/disattivazione ed eliminazione (sposta in .trash).

Pulsante di avvio/arresto con un clic fissato nell'angolo inferiore destro della finestra: riavvia DeepSeek Harness sul posto tramite un helper separato che rilancia l'esatta invocazione di avvio (protetto da loopback same-origin), con un'azione opzionale di solo arresto.

Consulente per la valutazione e il miglioramento delle PR: valuta la probabilità di successo delle PR in base ad anti-pattern storici e modelli di successo, con punteggio in un clic e suggerimenti pratici.

Toolkit vision per DeepSeek solo testo: strumento `vision` invocabile dal modello, adapter wrapper per deepseek/opencode-go (v4 flash/pro), quota Antigravity IDE (default, flash/pro) / qualsiasi VLM compatibile OpenAI / Gemini / canali locali Ollama, memoria delle evidenze con reidratazione da compattazione, cache dell’hash dei contenuti e un pannello client bilingue.

Test di accettazione browser indipendenti per i deliverable degli agenti: in input una specifica JSON, in output un verdetto da Chromium reale.

Oscura i pattern di segreti supportati dalla copia esportata di `session-telemetry/record` prima che raggiunga i backend di telemetria configurati.

Personal dsh preferences: a tunable companion to the official compaction engine (threshold, retention, summarization model/reasoning, turn-end compaction), llama.cpp model import for the Models page, and tab/notification helpers (status-light favicon, bro

DSH plugin: counts completed compactions per session and raises a floating reminder — with bilingual UI — only after repeated compactions plus heavy extra token spend.

Windowed, reversible context management for DeepSeek Harness; replaces Basic compaction.

Deterministic project actions shared by people and agents: one actions.json, a right-sidebar run panel, and agent tools.

DSH Web plugin that injects an SVG (feTurbulence + feDisplacementMap) refraction filter and applies it through backdrop-filter to panels, bubbles and the input card, and bundles the orca-scene skin with theme tokens plus light and dark backgrounds.

Compresses older DSH context into a structured handoff, retains recent messages, and keeps session history searchable.

Animated DSH Web pet with model-adaptive output pacing, tool-state reactions, replaceable character packs, and a bundled skill for creating or extending pet animations.

Cross-session experience memory for DSH. Lessons are evidence-graded, injected when the turn is about the same thing, and shown just before an action only when the record declares which call it applies to. A rule that holds whatever the turn is about — "always answer in Chinese" — can be marked **standing** and is then carried every turn: measured on 24 isolated real-model turns, a standing rule is followed 6/6 against 0/18 for the controls (empty store, same record unmarked, unrelated standing record). A guess still cannot become always-on: a standing record needs a verified passage like any other. What nobody looks up is retired, reversibly. Five tools, seven slash commands, zero runtime dependencies.

Simulated life context for DeepSeek Harness: every turn injects the past 24h of simulated life events from the workspace .life/ directory (sliding-window filter, session-level deduplication with delta updates, world summary) as a collapsed context notice, and the life_react tool lets the agent record feelings, thoughts and actions back into its life log for the next round of world evolution — turning a DSH agent into a roleplay character with a persistent daily life.

Compaction fallback for compaction-basic: when its LLM summary throws, comes back empty or cannot fit the window, the span is compiled deterministically with a seq pointer on every line (vendored dsh-compaction-instant compiler). Adds recall and search tools, /recall and /context commands, an idle trigger at 45% of the window and an on-disk span archive. The engine mounts as one row in the agent preset (docs/dsh-integration.md).

Model-autonomous context compaction for DeepSeek Harness: a mind_map tool surveys the agent context surface (stable seq ids, roles, token weights, clear-boundary markers, the clear-mind playbook), a clear_mind tool replaces a chosen conversation-history span with a model-written checkpoint through the native compaction transaction, call and result pairs self-collapse into one-line tombstones, and a proactive reminder nudges the model when the context grows long or a turn runs too many steps.

Model-callable compact_agents tool that force-compacts context in every live session (main session, ordinary sub-agents and AgentTeams members alike), queues busy targets and compacts them when the turn ends, reports the shadowed node count and estimated tokens per target, and ships a browser form for the trigger ratio, retained ratio, controlled-phase output budget, auto-continue budget and an end-of-turn pre-compaction ratio that compacts near the trigger line once a turn ends, so a reply no longer opens with a wait for the summarizer.

Distills session work into background observations and reflections to preserve context across compaction.