@acontext/openclaw

Acontext skill memory plugin for OpenClaw. Your agent learns from conversations, distills reusable skills as Markdown files, and syncs them to OpenClaw's native skill directory.

What it does

1. Auto-Capture — Stores each agent turn to an Acontext session, triggering automatic task extraction
2. Skill Sync — Downloads learned skills to ~/.openclaw/skills/ for native loading by OpenClaw
3. Auto-Learn — Triggers Learning Space skill distillation when enough conversation accumulates
4. Agent Tools — 3 tools for explicit skill/memory operations during conversations

Installation

openclaw plugins install @acontext/openclaw

Setup

Get an API key from dash.acontext.io and set it in your environment:

export ACONTEXT_API_KEY=sk-ac-your-api-key

Add to your openclaw.json:

{
  plugins: {
    // Select Acontext as the active memory plugin
    slots: {
      memory: "acontext"
    },
    entries: {
      "acontext": {
        enabled: true,
        config: {
          "apiKey": "${ACONTEXT_API_KEY}",
          "userId": "your-user-id"
        }
      }
    }
  }
}

Note: OpenClaw only loads one memory plugin at a time. Setting plugins.slots.memory to "acontext" replaces the default (memory-core). To switch back, set it to "memory-core" or "none".

Restart the gateway:

openclaw gateway

Configuration

| Key | Type | Default | Description | |-----|------|---------|-------------| | apiKey | string | — | Required. Acontext API key (supports ${ACONTEXT_API_KEY}) | | baseUrl | string | https://api.acontext.app/api/v1 | Acontext API base URL | | userId | string | "default" | Scope sessions per user | | learningSpaceId | string | auto-created | Explicit Learning Space ID | | skillsDir | string | ~/.openclaw/skills | Directory where skills are synced for native loading | | autoCapture | boolean | true | Store messages after each turn | | autoLearn | boolean | true | Trigger skill distillation after sessions | | minTurnsForLearn | number | 4 | Minimum conversation turns before triggering auto-learn |

Agent Tools

| Tool | Description | |------|-------------| | acontext_search_skills | Search through skill files by keyword | | acontext_session_history | Get task summaries from recent past sessions | | acontext_learn_now | Trigger skill learning from the current session |

CLI Commands

# List learned skills
openclaw acontext skills

# Show memory statistics
openclaw acontext stats

How it works

Capture → Extract → Learn → Sync

Session 1: User talks to agent
  └→ Messages stored to Acontext session (incremental capture)
  └→ CORE extracts structured tasks
  └→ Learning Space distills skills as Markdown

  ── Compaction or reset occurs ──
  └→ Pending messages flushed & learning triggered
  └→ Message cursor resets for next capture

Session 2: User returns
  └→ Skills synced to ~/.openclaw/skills/ (native loading)
  └→ OpenClaw auto-loads skill files
  └→ New messages captured (incremental capture)
  └→ Skills updated (auto-learn)

Note: Message capture is incremental — the plugin tracks a cursor so only new messages are stored on each agent_end. When compaction or reset rewrites the transcript, the cursor resets automatically.

Skill sync

Skills are synced as Markdown files to OpenClaw's native skill directory (~/.openclaw/skills/ by default). Only changed skills are re-downloaded using server-side updated_at timestamps. Sync happens on service start, before each agent turn (if stale), and after learning.

How it differs from fact-based memory

Unlike plugins that extract discrete facts via LLM, Acontext stores full conversations and distills them into human-readable, editable Markdown skill files. You can inspect, modify, and share these files directly.

License

Apache-2.0