DeepSeek Harness Plugin Hub

Publish and manage complete Harness Profiles. Discover Plugins for your next setup.

Explore

PluginsPresetsDocsNews

Community

Publish a pluginContactReport an issue

Resources

Plugin Hub on GitHubDeepSeek HarnessSystem statusPrivacy notice
© 2026 DeepSeek Harness Plugin HubPowered byPaxTech

Independent and unofficial. Not affiliated with, authorized by, or endorsed by DeepSeek.

Plugin Clm — DSH Plugin for DeepSeek Harness
DeepSeek Harness Plugin Hub
ProfilesPluginsCategoriesNewsDocsSign inManage Profiles
ProfilesPluginsCategoriesNewsDocsSign in
← Plugins
P

dsh-plugin-clm

Plugin Clm

Context Language Models (CLM) for DeepSeek Harness: Active editable context window mirrored to disk for pruning and token optimization.

The plugin will be installed here. Keep web if you are unsure.

npx -y @deepseek-ai/dsh plugin --profile web add github:maxeone/dsh-plugin-clm#7dad05fea89fcc54825840fd2243a815762d24f8
READMECompatibilityVersions

Compatibility and provenance

Plugin Clm is published as dsh-plugin-clm and currently resolves to version 1.0.2. The Hub verifies its manifest and preserves the exact installation source for reproducible installs.

DSH compatibility
*
Runtime surfaces
any
Release source
github
Registry updated
10/12/2026

Versions

1.0.2stable
10/12/2026
1.0.1stable
10/12/2026

Related plugins

Loading related plugins…

Latest
1.0.2
DSH
*
HMR
Process restart
Tree shaking
Safe tree shaking not declared
Unpacked size
Unavailable
Files
Unavailable
Surface
any
License
MIT
Source
github
GitHub
★ 0
Weekly downloads
0
Last push
10/12/2026
View source ↗
README badge

Click the badge to copy Markdown for your README.

Do you maintain this Plugin?Claim benefit · Priority security scan

Verify the GitHub repository declared in package.json to manage this listing. After you claim it, Hub will prioritize a security scan of the current version and publish the result when it passes.

Claim this Plugin →
Report an issue

Related plugins

More verified plugins in memory-context.

Hindsight Coding Agents@vectorize-io/hindsight-coding-agentsReflect-only Hindsight long-term memory for coding agents (harness-pluggable: opencode, …), with automatic background ingestion (no setup CLI).Ab Memorydsh-ab-memory--- description: "The dsh-ab-memory package: a cross-session, file-backed agent memory with five tools (remember / recall / forget / load_skill / distill) and an always-on light injection." kind: "package-reference" ---Contextdsh-contextA DeepSeek Harness plugin for context insight and management, with context dashboard and context command, for understanding how the context is made of, and how it evolves.Memsearch Dsh@zilliz/memsearch-dshMemSearch plugin for DeepSeek Harness: shared markdown memory across agents, with capture, pre-step context injection, memory-recall skill, and a skill-candidate review panel.

README

🧠 dsh-plugin-clm (Context Language Models for DeepSeek Harness)

Active editable context window mirrored to disk for DeepSeek Harness (DSH). Inspired by Meta CLM (Context Language Models) & pi-clm.


🇪🇸 Descripción en Español

dsh-plugin-clm transforma la ventana de contexto estática y puramente acumulativa de DeepSeek Harness en un documento de contexto vivo y editable en disco (.dsh/clm/LIVE_CONTEXT.md).

En las interacciones tradicionales con LLMs, el historial de conversación crece sin control, acumulando volcados gigantescos de terminal, salidas de búsqueda obsoletas y caminos sin salida que desperdician tokens y degradan el razonamiento del modelo. Con CLM, el agente y el usuario pueden podar, resumir y mantener activamente el contexto de la conversación.

✨ Características Principales

  • 🔄 Sincronización Bidireccional en Tiempo Real: Refleja la memoria activa del turno en .dsh/clm/LIVE_CONTEXT.md.
  • ✂️ Poda Activa de Contexto: El modelo puede comprimir salidas extensas de herramientas (pwsh, curl, grep) en notas compactas de 1 o 2 líneas.
  • 🎛️ Toggle Interactivo en la Interfaz Web: Botón [ 🧠 CLM ON 🟢 / CLM OFF ⚪ ] directamente integrado en la barra de herramientas del chat, junto a los permisos y el modo Plan.
  • 🛡️ Overflow Guard: Previene desbordes accidentales de la ventana de contexto asegurando un presupuesto estricto de tokens (budgetTokens: 64000, reserveTokens: 4096).
  • ⚡ Persistencia por Conversación: Recuerda el estado encendido/apagado para cada sesión de chat individualmente.

🇬🇧 English Description

dsh-plugin-clm brings Meta's Context Language Models (CLM) paradigm to DeepSeek Harness. Instead of an append-only monolithic context, your conversation history is mirrored into a structured, editable Markdown file (.dsh/clm/LIVE_CONTEXT.md).

✨ Key Highlights

  • 🔄 Live Disk Mirror: The conversation context is projected into an editable document before each generation turn.
  • ✂️ Active Pruning: Both user and agent can prune bloated tool outputs, dead ends, or add persistent project notes.
  • 🎛️ Native UI Toggle: Integrated status chip [ 🧠 CLM ON / OFF ] right in the conversation composer bar.
  • 🛡️ Overflow Guard: Automatic safeguard to keep generation within token budgets.

🚀 Instalación / Installation

En tu perfil de DeepSeek Harness (~/.dsh/profiles/desktop/package.json):

{
  "dependencies": {
    "dsh-plugin-clm": "^1.0.0"
  },
  "dsh": {
    "profile": {
      "bundles": [
        "dsh-plugin-clm"
      ]
    }
  }
}

O instala directamente vía CLI:

dsh plugin add dsh-plugin-clm

🛠️ Comandos Slash / Slash Commands

ComandoAcción
/clm onActiva la memoria viva editable en la conversación actual
/clm offDesactiva CLM y regresa al flujo acumulativo tradicional
/clm statusMuestra el tamaño estimado en tokens y la revisión del espejo
/clm resetReinicia y regenera el archivo LIVE_CONTEXT.md limpio

📄 Estructura del Archivo Espejo (LIVE_CONTEXT.md)

# CONTEXT LANGUAGE MODEL (CLM) - MEMORIA VIVA EDITABLE

[[CTX_NOTES]]
- Objetivo: Refactorizar arquitectura de autenticación.
- Decisiones: Mantener tokens JWT en memoria protegida.

[[CTX_TURN document=sess-1 index=1 role=user]]
Hola, ayúdame a revisar los logs del servidor.

[[CTX_TURN document=sess-1 index=2 role=assistant]]
Analizando registros:
[Salida de logs comprimida a los 3 errores críticos detectados]

👤 Autor & Licencia

Desarrollado por Maximiliano Roldán (@maxeone).
Distribuido bajo licencia MIT.