context-compactor

Token-based context compaction for local models (MLX, llama.cpp, Ollama) that don't report context limits.

7 stars

Best use case

context-compactor is best used when you need a repeatable AI agent workflow instead of a one-off prompt.

Token-based context compaction for local models (MLX, llama.cpp, Ollama) that don't report context limits.

Teams using context-compactor should expect a more consistent output, faster repeated execution, less prompt rewriting.

When to use this skill

  • You want a reusable workflow that can be run more than once with consistent structure.

When not to use this skill

  • You only need a quick one-off answer and do not need a reusable workflow.
  • You cannot install or maintain the underlying files, dependencies, or repository context.

Installation

Claude Code / Cursor / Codex

$curl -o ~/.claude/skills/context-compactor/SKILL.md --create-dirs "https://raw.githubusercontent.com/Demerzels-lab/elsamultiskillagent/main/public/skills/emberdesire/context-compactor/SKILL.md"

Manual Installation

  1. Download SKILL.md from GitHub
  2. Place it in .claude/skills/context-compactor/SKILL.md inside your project
  3. Restart your AI agent — it will auto-discover the skill

How context-compactor Compares

Feature / Agentcontext-compactorStandard Approach
Platform SupportNot specifiedLimited / Varies
Context Awareness High Baseline
Installation ComplexityUnknownN/A

Frequently Asked Questions

What does this skill do?

Token-based context compaction for local models (MLX, llama.cpp, Ollama) that don't report context limits.

Where can I find the source code?

You can find the source code on GitHub using the link provided at the top of the page.

SKILL.md Source

# context-compactor

Token-based context compaction for local models (MLX, llama.cpp, Ollama) that don't report context limits.

## Install

```
npx clawhub@latest install context-compactor
```

Related Skills

smart-context

7
from Demerzels-lab/elsamultiskillagent

Token-efficient agent behavior — response sizing, context pruning, tool efficiency, and delegation.

project-context-sync

7
from Demerzels-lab/elsamultiskillagent

Keep a living project state document updated after each commit, so any agent (or future session) can instantly understand where things stand.

auto-context-manager

7
from Demerzels-lab/elsamultiskillagent

AI-powered automatic project context management.

context-builder

7
from Demerzels-lab/elsamultiskillagent

Generate LLM-optimized codebase context from any directory using context-builder CLI.

contextkeeper

7
from Demerzels-lab/elsamultiskillagent

ContextKeeper — Safe project state tracking for AI agents.

context-viz

7
from Demerzels-lab/elsamultiskillagent

Visualize the current context window usage — token estimates per component (system prompt, tools, workspace files.

telegram-context

7
from Demerzels-lab/elsamultiskillagent

Toggle-enabled skill that fetches Telegram message history at session start for conversational continuity.

multi-chat-context-manager

7
from Demerzels-lab/elsamultiskillagent

CLI tool to store and retrieve conversation contexts per channel/user.

context-gatekeeper

7
from Demerzels-lab/elsamultiskillagent

Keeps the conversation token-friendly by summarizing recent exchanges, surfacing pending actions, and delivering.

slack-context-memory

7
from Demerzels-lab/elsamultiskillagent

Conversation summarization and context compaction for Slack channels. Reduces context window usage by 70-99% while preserving key information through semantic summaries.

context-onboarding

7
from Demerzels-lab/elsamultiskillagent

Provide new contributors and agents with a concise tour of the workspace identity files (SOUL.md, USER.md, AGENTS.md, TOOLS.md) plus onboarding tips. Use when a newcomer needs context or when you want to double-check how this workspace is configured.

context-anchor

7
from Demerzels-lab/elsamultiskillagent

Recover from context compaction by scanning memory files and surfacing where you left off. Use when waking up fresh, after compaction, or when you feel lost about what you were doing.