context-management

dsh-nested-followups

sluminositys

Branch from any answer and keep branching at any depth. Recursively isolated follow-ups that leave the main DSH task untouched.

dsh-token-optimizer

Zoria-Lind

Layered token-optimization pipeline for DeepSeek Harness: output ladder, MCP lazy loading,compaction driver, cache-hit reporting. Built on real DSH plugin APIs; ~40-60% input saved in long sessions.

taskfold

yindf

Context folding for DeepSeek Harness (DSH): wrap work in named tasks, fold finished spans into short summaries, recall originals on demand.

fast-compaction-dsh

kolawong

Verdict-based context compaction for DeepSeek Harness — replaces lossy LLM summaries with fast keep/truncate/drop decisions from jev-latest; everything kept stays verbatim. Port of tamaratran/fast-jev-compaction.

dsh-dcp

chiro2001

DeepSeek Harness 动态上下文管理插件（Dynamic Context Pruning for dsh），对标 opencode-dcp

dsh-context-checkpoint

Leafstory

为解决 DeepSeek V4.1 Flash 在 DSH 长上下文下丢失结论 / 幻觉升高而做的自动总结-压缩-注入插件：达限 → 总结落盘 → 自动压缩 → 压缩后注入回上下文开头（含 skill 与一键安装器）

context-guardian

LuminariSoftwares

Compaction for local models that actually fires and never bricks the session. OpenAI-compatible proxy + DeepSeek Harness engine.

tool-guardian

LuminariSoftwares

Your MCP servers for ~300 tokens instead of ~28,000. MCP server + native DeepSeek Harness bundle.

rcs-context

ScreamingMaggot

把长对话的重复计费砍掉一半，并给模型配一名只读调查员：先核实，后动手。面向 DeepSeek Harness 的上下文压缩与外置推理插件（MIT）。

dsh-tree-task-flow

viyiviyi

树形任务流 · DeepSeek Harness 插件：把多步任务组织成目标→任务→子任务三级树，节点完成时由 AI 提交结果，插件随即把该节点的执行过程从上下文里折叠掉。

dsh-jev-context

YIRC99

面向 DeepSeek Harness 的 JEV 上下文管理插件：按相关度裁剪与召回历史片段，可视化追踪每轮决策和 token 变化。