Model-driven context management (Active Context Pruning / ACP) for the DeepSeek Harness — the model decides when and what to compress. Ported from billion-context-pi (ranxianglei); acp-kernel reused verbatim. CompactionEngine backend with compress/decompress/search_context/acp_status tools.
DSH 插件:直接复用 Codex CLI 本地登录订阅凭证,在 DeepSeek Harness 中使用 ChatGPT 订阅模型,无需 API Key | DSH plugin: reuse your Codex CLI local subscription login to use ChatGPT subscription models in DeepSeek Harness, no API key required
Silent vision bridge for DeepSeek Harness: route chat images to a fixed vision model, preserve UI originals, and reuse observations across compaction and restarts.
Project-level change routing and reuse governance for AI coding agents. Helps Codex and Claude Code find the right capability boundary, avoid duplicate implementations, respect write constraints, and calibrate routing decisions for large full-stack repositories.
Use DeepSeek Harness (DSH) directly inside the VS Code sidebar by embedding its web UI. Reuses an already-running `dsh web` server, or launches one for you.
Desktop launcher shell for DeepSeek Harness (dsh): a tray-resident Electron wrapper that starts or reuses the local dsh web server and opens it in a native window
Reuse your own logged-in Chrome in dsh (DeepSeek Harness): scrape login-walled sites, fill forms, and click through pages with native wc3_* tools — no re-login, no anti-bot friction.
Pre-write reuse firewall for DeepSeek Harness: before the agent writes a new helper/service, surface the existing implementations that already cover that intent. Deterministic retrieval (no LLM) via the Auto_code_audit capability channel.
