114 results
Automatic per-task model hot-switching for DeepSeek Harness (dsh): image-aware tasks route to the vision model automatically, every other task keeps your default model. Zero extra tokens, no context disturbance.
Probe local model servers (llama.cpp / Ollama) for the real runtime context window and keep llm-pi-ai's contextWindow in sync automatically — tighter before overflow, wider when the server grows. Zero-config profile plugin for DeepSeek Harness.
会话注入门控：给 DeepSeek Harness (dsh) 加一个会话级上下文裁剪闸门，按会话取消/恢复插件贡献的 skill、tool、系统提示词段落——3 条影子通道、5 条本地路由、零依赖零配置。Per-session injection gate for DeepSeek Harness (dsh): trim skills, tools and prompt sections per session. Zero-config, zero dependencies.
Keep native vision, drop historical images before they hit relay providers. Global DeepSeek Harness bundle for every agent.
Multi-engine web search for DeepSeek Harness: an aggregated ctx.web search provider (auto-routing + sequential fallback, free engines keep search alive) plus a web_search_pro model tool with per-query engine choice, fan-out, batch and verticals — Brave, B
Recover unfinished DSH conversations after restart, keep active goals running, and preserve questions waiting for your answer.
Session image payload governor for DeepSeek Harness: audit which images a session still ships, pick the ones to keep, and move the rest out of the model context.
DeepSeek Harness plugin: keep every model request inside an allowed set per provider, redirect or refuse the rest, and warn when a request would make a llama.cpp router reload models
Keep DSH tool results cache-prefix safe: truncate a result before it first enters the session surface, and never rewrite a tool result a provider request has already delivered.
GitHub tools for DeepSeek Harness: 18 agent-facing github_* tools plus a Web settings page that keeps the PAT in the DSH credential store.
Per-session auto-resume toggle for long-running goal sessions: mark a session, and after the harness restarts the agent is resumed automatically and its active goal keeps running.
DeepSeek Harness plugin that keeps rolling, integrity-verified backups of each session at turn boundaries, so any corruption costs at most the in-flight turn; one-command restore, even when dsh won't boot
dsh-plugin: quality-first topic memory for dsh — a topic keeps the starting question, conclusion, impact and dependencies, not the process; local-first git-traced OKF bundle, budgeted LLM-free injection
Keep long DSH chat sessions responsive: skip layout/paint for off-screen transcript rows (content-visibility).
Keep the Windows console on code page 65001 (UTF-8) for the dsh host and its bash tool commands, so output from Windows-native child processes stops being decoded as mojibake
Codex-style folding for the DeepSeek Harness web transcript: a running turn keeps only its last few actions visible, folded behind one row that opens the whole turn.
dsh-3301 — LAN gate for a DeepSeek Harness (DSH) Web GUI: keeps DSH loopback-only, exposes a password-protected entry (HTTP + WebSocket) with a settings card, and stores the password only as a scrypt verifier. Ships as a DSH plugin and as a standalone CLI.
Keeps the browser tab in step with the DeepSeek Harness workspace: a status prefix on the page title and a count badge on the favicon, for sessions that are running, that finished while you were elsewhere, or that are waiting on your answer.