dsh-whale
1Lyn-en
DeepSeek Harness minimal-reply plugin, offering six brevity modes and a CyberUI theme, saving 60–75% of output Token|A DSH plugin with six brevity modes and a CyberUI theme.
GITHUB TOPIC
12projects include this topic
The "token-saving" topic on GitHub groups 12 open-source projects in the DeepSeek Harness (DSH) ecosystem, led by dsh-whale with 2 GitHub stars. dsh-whale — DeepSeek Harness minimal-reply plugin, offering six brevity modes and a CyberUI theme, saving 60–75% of output Token|A DSH plugin with six brevity modes and a CyberUI theme. Every project here is indexed by DSH Universe with live GitHub data — stars, activity and install status — so you can compare and install directly.
{count} projects
Exact GitHub Topic match
1Lyn-en
DeepSeek Harness minimal-reply plugin, offering six brevity modes and a CyberUI theme, saving 60–75% of output Token|A DSH plugin with six brevity modes and a CyberUI theme.
Mungbean-Cake
DeepSeek peak/off-peak reminder plugin: live Beijing-time announcements, advance alerts before switching, token-saving tips, freely draggable
anyi-520
DeepSeek Harness 插件:把 skill 与 MCP 从默认工具定义里抽出来,改成按需热加载的懒工具箱(skill / mcp 分开列出,一句话说明用途,用时注入、用完卸载)
chimneyQ
DeepSeek Harness bundle: native ctk_* sandbox tools, a local FTS5 knowledge base, and a routing gate that keeps bulk output out of the conversation. No third-party runtime dependencies.
mili1360
DSH 会话越长越慢、每轮越贵?把最老的历史折叠成一条带索引的消息,全文归档到本地、随时能沿路径找回来,不调用模型写摘要。DeepSeek Harness plugin: folds the oldest history span into one indexed message backed by a lossless local archive.
MythGreenLand
Delegate mechanical text work (translate/summarize/extract/classify) to a local cheap model endpoint, keeping large text out of the main model context.
pythonshiyi
DeepSeek Harness agent preset: full Standard tools plus an extreme compact-output expert prompt
xiao-hans
基于 RTK、专为 DSH 打造的薄层适配插件:让 DSH agent 的 pwsh 命令自动走 RTK(Rust Token Killer)压缩输出、节省 LLM token。A thin DSH adapter built on RTK that auto-routes DSH agents' pwsh commands through RTK to cut token usage.
yzsp-studio
🔥 Cut 91% of tokens, lose zero capability | 省 91% token · 能力零损失 — Scope DSH skill catalogs & MCP tool schemas per workspace. Measured: 36,870 → 11,658 tokens/step (−68%). Nothing deleted. | 按工作区裁剪技能目录与工具 schema,每步降 68%,一个字都不删。📮 wow.vip@foxmail.com
Nolitor
TokenEconomizer for DeepSeek Harness — turn-based output tiering, tool-result dedup, KV-cache-safe prompt slot, loop breaker with thinking watchdog, prior-session recall, and the microsoft/TokenEconomizer engine bridge. Package: @nolit/dsh-tokeneconomizer.
WODE25500
DSH token-saving toolkit: proactively slim large text/JSON/CSV before they enter context (diet_text / diet_json / diet_csv / diet_estimate). Zero-dependency pure functions.
YansongG-HKU
Native DeepSeek Harness (DSH) plugin: reversible incremental context compaction with model-facing tools (status/search/restore/compact) and a Web panel — save tokens on long sessions without dumbing the model down.