managed-agents
Local-first, self-hosted AI agent runtime with Claude Managed Agents-style APIs, sandboxed sessions, memory, tools, audit, replay, and a local Console.
31 results
Local-first, self-hosted AI agent runtime with Claude Managed Agents-style APIs, sandboxed sessions, memory, tools, audit, replay, and a local Console.
DeepSeek Harness 移动端适配与安全访问插件,支持局域网、远程连接、Android App 和手机浏览器。
dsh-passwords: a server-grade gateway that turns DeepSeek Harness into a multi-tenant platform — remote access + automatic HTTPS, per-subuser permissions & quotas, sandbox enforcement, first-run setup, SQLite auth with at-rest encryption, rate-limit and audit log (bilingual zh/en UI)
DeepSeek Harness plugin for remote access: a token-gated reverse proxy keeps settings, credentials, and file access working over public tunnels and on other devices instead of returning 403. Per-device sessions.
Multi-provider Web Search & Fetch for DeepSeek Harness — 8 deeply adapted providers, SearchHints, resilient fallback, and native X / Xiaohongshu retrieval
Authenticated remote access for a DeepSeek Harness web profile: TLS, QR/passcode device pairing, password sign-in, and per-device revocation in front of an untouched loopback harness.
dsh plugin: back the native web_search / web_fetch tools with your self-hosted SearXNG instance - keyless, private, no third-party search vendor.
dsh-yolo-mode —— DeepSeek Harness 双面包插件:当会话处于可写沙箱模式且审批策略为 ask 时,用大模型自动裁决沙箱升权申请,支持内置预设与自定义权限层级,并提供宿主 settings + 自发布设置桥(/yolo-mode)与 Web 客户端 UI。
Password-gated public access for the Web GUI via Cloudflare Tunnel: a loopback password gate (login page + HMAC cookie) in front of the webserver, published by spawning cloudflared in quick or named-tunnel mode - tells the shell and the model the public URL
Native DeepSeek Harness plugin marketplace for discovering, filtering, installing, and managing community DSH plugins.
SearXNG-backed web search provider for the DeepSeek Harness web capability seam (ctx.web) — self-hosted metasearch, no API key, no model turn per search
企业 AI 资源统一管理平台 —— 基于 DeepSeek Harness「一切皆插件」架构
Ollama native protocol (/api/chat) provider for DeepSeek Harness: llm-ollama settings namespace, serving LlmAdapter, /api/tags model discovery, and a whole-page Ollama settings section.
DeepSeek Harness tools for self-hosted Gitea and Forgejo instances.
Cross-session long-term memory for AI agents — local Go service (~50MB RAM, SQLite-vec, verify-augmented) with bridges for DeepSeek Harness, Claude Code, Codex, Hermes, OpenClaw, opencode, pi, Reasonix.
DeepSeek Harness mobile pairing remote plugin: E2EE companion over a dual-plane allowlisted RPC (LAN / Tailscale / optional Cloudflare Quick Tunnel or self-hosted rendezvous).
Let a local DeepSeek Harness Agent use BailingHub to query data, update records, and run permitted actions in your online store, SaaS, or other admin system.
DeepSeek Harness bundle providing HTTP Basic Authentication, a bilingual web login page, a Web UI settings card, remote IP access, and a Web Cryptography UUID polyfill for self-hosted server deployments.
One-click updates for DeepSeek Harness (DSH). Community plugin — incremental npm tarballs, backup, rollback. Not official DeepSeek software.
dsh plugin: back the native web_search / web_fetch tools with your self-hosted SearXNG instance - keyless, private, no third-party search vendor.
DeepSeek Harness relay: mirror the home dsh web UI to the public internet through an outbound wire trunk (cloud relay + out-of-tree host plugin).
Mobile companion layout for the dsh web GUI: phone-width drawer sidebar, overlay details pane, touch-sized composer and safe-area handling
Mem0 persistent memory for the DeepSeek Harness web profile — automatic recall injection, tidal-coalesced memory writes, and mem0_search/add/update/delete tools against a self-hosted Mem0 server (X-API-Key). Zero-intrusion bundle plugin; no dsh source changes.
OpenAI-completions-compatible adapter for custom gateways (vLLM / LM Studio / self-hosted proxies): always role:"system", enable_thinking / chat_template_kwargs / reasoning_effort driven by the llm-pi-ai model config, and a </think> split on the receive side — no pi-ai behavior guessing.