Agent collaboration · DeepSeek Harness plugins — Page 123

ai-bridge suite dashboard: one settings tab aggregating topology/observe/router/orchestrator/mcp-bridge live state

MCP multi-server orchestration for DeepSeek Harness: aggregated server config via hot-reload settings, dynamic spawn/dispose of dsh-mcp-client instances, and a Bridge tab for status

Non-invasive observability timeline + statistics for DeepSeek Harness Web UI (tool calls, LLM streams, live config); data foundation for the ai-bridge orchestration suite

Multi-agent orchestration for DeepSeek Harness: five dispatch modes (parallel/sequential/select/cascade/merge) with retry, executed through the native subagent seam

Explainable Bayesian subagent-provider routing for DeepSeek Harness: scores providers from observed run history and ranks them per task (7-dimension core)

Live Cordis plugin dependency topology graph for DeepSeek Harness Web UI

验证 DSH 插件的 CLI：一条命令跑 mock-llm 完整 agent 循环，检查 waterfall 链完整性与零副作用，产出验证报告

Simulator State as Code for iOS developers, CI pipelines, MCP clients, and DeepSeek Harness

Fail-closed semantic policy guard for secure DeepSeek Harness tool execution.

DeepSeek Harness guardrail that limits how many files an AI coding agent can edit per turn

Persistent dual-face Design Studio plugin for the DSH harness: the design_studio agent tool (design systems under temp_design_folder/, identity presets, OpenRouter vision review, and a real harness-subagent design agent that lists, reads and edits the design files itself), the /design-studio live-preview route + JSON API, a design-brief prompt section, and the built-in Design Studio web UI (conversation tab, agent chat, settings pages). Install once — the tab appears on every launch. Zero runtime dependencies (plain JS).

Persona engine for AI coding harnesses (DeepSeek Harness; the ZCode adapter is retired as of 2026-09-19 and kept for history only): self-name per model, self-image (appearance) and reply tone per model (both opt-in), user address, stance, character, per-item work contracts, thinking-chain language, a long-term memory inbox where AI entries only take effect after an explicit human confirm line (enforced in code), and a shareable persona-preset library that imports SillyTavern character cards (v1/v2 JSON; unmappable fields are reported, never faked). One shared config.json, zero runtime dependencies, never touches the network.

Debate orchestrator for DeepSeek Harness: a dsh profile bundle shipping the debate_open/debate_round/debate_verdict tools and their Web card

Shopify plugin for the DeepSeek Harness: 210+ model-facing tools over the Shopify Admin REST and GraphQL APIs (products, orders, customers, inventory, fulfillments, discounts, content, webhooks, themes, marketing, billing and more), with Admin API access-token auth via env-var references.

DSH plugin for the llm-wiki knowledge base: OKF v0.2-conformant markdown bundle with progressive disclosure. Registers 13 tools: wiki_list/search/get/create/update/validate/lint/ingest/deprecate/rules/help/dirs/use. Named multi-bundle directories with session/global switching. Self-contained (vendor-core embedded).

Harness-native J-Space cognition preset with cache-stable scoped prompt control.

Generate a single suggested next prompt for the user after each completed agent turn, surfaced as ghost text in the web composer

A safety harness plugin for DeepSeek Harness (DSH): protected-path enforcement, trash-based safe_delete with undo, last-known-good composition snapshots, pre-restart composition checks, and an audit journal.

Agent Plugins 1.0.0 adapter for DeepSeek Harness: discovers skills/ and mcp.json from agent-plugin stores, registers skills and MCP servers, and manages them through a CLI and the settings panel.

Cross-session tools and Web @session mentions for DeepSeek Harness.

The authority layer for agentic AI, as a DeepSeek Harness plugin. Governs every tool call against your business rules via the TapPass /v1/govern policy decision point.

Bidirectional bridge between Hermes Agent and DeepSeek Harness. Hermes dispatches focused tasks via JSON-RPC over POST /mcp/collab (one-shot or continuable, nonce-bound amend, secret-bound consult reply); DSH lets you continue Hermes sessions as native DSH sessions in the sidebar.

OpenTelemetry traces, metrics, and logs for DeepSeek Harness, written straight into GreptimeDB.

Workspace-free temporary conversations for DeepSeek Harness