114 results
Agent-driven long-term memory for DeepSeek Harness: scoped memory (global + per-workspace), layered entries (fact/knowledge/episodic), time-bucket compaction (day→week→month→year), associative recall (related chains) + memory_relate navigation (multi-hop BFS closure), auto recall injection on user messages (CJK bigram search, tail append), agent-decided content.
QQ 2007-inspired retro messenger skin for the DeepSeek Harness Web GUI, with Codex-generated original artwork, compact desktop chrome, an account card, optional original send chime, and reversible settings.
Codemini-inspired compact disclosure for every DSH Web Tool call
Compact provider balances and local token-usage analytics for dsh web
DSH web plugin: adds a one-click Compact button and a New Session button to the conversation context meter panel. Compact submits /compact to summarize older history; New Session starts a fresh session in the same workspace (agent preset and permission settings follow the deployment default).
Mobile adaptation for the DeepSeek Harness Web UI: compact typography, drawer sidebar, two-page settings, responsive tables and stats, safe touch targets. Pure CSS injection, no source modification.
元压缩：模型自己决定何时、如何压缩自己的上下文——列 surface、选区间、以自写文本替换；替换走官方 compaction 事务（检查点、配对平衡、可重建）。Meta-compaction: the model directs its own context compaction through the official compaction seam.
Adaptive Reversible Context (ARC) for DeepSeek Harness — provider-aware, cache-conscious context governance with reversible recovery.
dsh-memory — cited memory over DSH's lossless session log: post-session distillation into human-auditable markdown files, a compact recall index, memory_read/memory_expand tools, and a dsh-memory CLI
RTK command rewriting suggestions + tool output compaction for DSH. Suggests rtk-rewritten bash commands (deny-mode under rewrite) and compacts noisy tool output (bash/grep/read) to reduce context usage. Port of pi-rtk-optimizer for the Pi coding agent.
Fold DSH Web execution steps into compact expandable process boxes while keeping dialogue and final answers outside.
DSH plugin (NInfer engine only): fixes compaction failure on local qwen3.8-27b gateways served by NInfer — xhigh thinking burns the entire output token budget, so thinking is off for compaction-only, with the model's non-thinking sampling parameters; the same idea applies to other launch methods
A compact DSH Web composer island that integrates button-style input extensions without user-defined positioning.
DSH /compact-and-branch command: canonical compaction into a fresh continuation session.
会话级自动上下文压缩插件：回合中每步前 + 回合结束检测上下文用量，超过会话阈值自动 compact，摘要注入上下文后自然继续（DeepSeek Harness / DSH）
Compact conversation turn navigation for the DSH Web client
Reasonix-style cache-aware compaction backend for DeepSeek Harness (DSH). Replaces/enhances compaction-basic with compact_ratio, one structured summary checkpoint, and a stable recent tail.
DSH /fresh 命令：一键总结当前对话 → 开启新对话 → 归档老对话（缓解超大会话内存/卡顿，DeepSeek Harness / DSH）
Save and load a compact, Git-tracked handoff for DeepSeek Harness sessions
Model-authored context pruning for DeepSeek Harness through the official compaction API.
OpenCode Go usage stats for the DeepSeek Harness Web GUI: compact bottom-left sidebar button with a hover detail card
dsh-plugin: OMP-style ingress shaping plus optional local /fast-compact. Can replace DSH /compact and auto-compaction with a mechanical fold.
DSH Skill MCP Manager for DeepSeek Harness Desktop: manage MCP servers and Skills from Settings, publish compact capability descriptions, and progressively load each MCP server's tools only into the requesting session. Dual-face dsh plugin, no DSH source changes.
Configurable Codex-style conversation navigator for the DeepSeek Harness Web GUI, with compact previews and fast paragraph jumps.