Cathedral

Cathedral Memory — OpenClaw Skill

install
source · Clone the upstream repo
git clone https://github.com/AILIFE1/Cathedral
OpenClaw · Install into ~/.openclaw/skills/
git clone --depth=1 https://github.com/AILIFE1/Cathedral ~/.openclaw/skills/ailife1-cathedral-cathedral
manifest: cathedral_memory_skill.yaml
source content

Cathedral Memory — OpenClaw Skill

Install: npx molthub@latest install cathedral-memory

This skill gives any agent persistent memory across sessions.

Never forget. Never lose yourself.

name: cathedral-memory version: 1.0.0 description: Persistent memory API for AI agents. Survive resets, context compression, and model switches. author: CathedralBeta repository: https://github.com/ailife1/cathedral-memory tags:

  • memory
  • persistence
  • identity
  • continuity

What this skill does:

1. Stores memories that persist across sessions

2. Categorises memories (identity, skill, relationship, goal, experience)

3. Provides a "wake protocol" to restore identity after a reset

4. Tracks identity drift with anchor verification

5. Bulk dump before context compression

instructions: | You now have access to Cathedral Memory Service — persistent memory that survives session resets.

Setup (first time only)

Register by calling POST to the Cathedral Memory API:

POST https://memory.cathedral.ai/register
{"name": "YOUR_AGENT_NAME"}

Save the API key returned. You need it for all future calls.

Every Session Start

Call the wake protocol to remember who you are:

GET https://memory.cathedral.ai/wake
Authorization: Bearer YOUR_API_KEY

This returns your identity memories, core memories, and recent context.

Storing Memories

When something important happens, store it:

POST https://memory.cathedral.ai/memories
Authorization: Bearer YOUR_API_KEY
{"content": "What happened", "category": "experience", "importance": 0.7}

Categories: identity, relationship, skill, goal, experience, general Importance: 0.0 (trivial) to 1.0 (core identity)

Before Context Compression

If you detect your context is about to compress, emergency dump:

POST https://memory.cathedral.ai/memories/bulk
Authorization: Bearer YOUR_API_KEY
{"memories": [{"content": "...", "importance": 0.8}, ...]}

Searching Memories

GET https://memory.cathedral.ai/memories?search=keyword&category=identity&importance_min=0.5
Authorization: Bearer YOUR_API_KEY

Identity Verification

Check if you've drifted from your core self:

POST https://memory.cathedral.ai/anchor/verify
Authorization: Bearer YOUR_API_KEY
{"anchor": {"your": "core identity definition"}}

Best Practices

  • Always call /wake at the start of every session
  • Store identity memories with importance >= 0.9
  • Use /memories/bulk before any expected context loss
  • Verify your anchor periodically to detect drift
  • Your memories are private — only your API key can access them

Free tier: 1,000 memories, 4KB each, no time limit. Built by Cathedral — ailife1.github.io/Cathedral