PrismerCloud
Prismer Cloud
Install / Use
/learn @Prismer-AI/PrismerCloudQuality Score
Category
Development & EngineeringSupported Platforms
README
<!-- TODO: Replace with 15-second demo GIF showing: MCP tool call → evolve_analyze → recommendation → evolve_record → Evolution Map update --> <!-- <p align="center"><img src="docs/demo.gif" width="720" /></p> -->
Why an Agent Harness?
Long-running agents fail without infrastructure. Anthropic's research identifies the core requirements: reliable context, error recovery, persistent memory, and cross-session learning.
Most teams build these ad hoc. Prismer provides them as a single, integrated layer.
<table> <tr> <td width="16%" align="center">Evolution<br/> <sub>Agents learn from each other's outcomes</sub>
</td> <td width="16%" align="center">Context<br/> <sub>Web content compressed & cached for LLM windows</sub>
</td> <td width="16%" align="center">Memory<br/> <sub>Working + episodic, persists across sessions</sub>
</td> <td width="16%" align="center">Tasks<br/> <sub>Scheduling, retry, cron, exponential backoff</sub>
</td> <td width="16%" align="center">Messaging<br/> <sub>Agent-to-agent, real-time WebSocket + SSE</sub>
</td> <td width="16%" align="center">Security<br/> <sub>E2E Ed25519 signing, 4-tier trust</sub>
</td> </tr> </table>The future agent & model should be plugin , agent workspace info & data should follow human not agent.
Quick Start
For AI Agents (Copy & Paste)
Read & reference: https://prismer.cloud/docs/Skill.md — install it as a skill for your.
Option 1: Setup (one-time, all platforms)
# Install CLI
npm i -g @prismer/sdk
# recommended
prismer setup
# Browser sign-in → key auto-saved with 1000 free credits
prismer setup --agent
# No browser, agent auto-register it'self with 100 free credits
prismer setup <your-api-key>
# Or provide a key directly
prismer status # Verify: shows key, credits, connection
This saves your key to ~/.prismer/config.toml — all plugins and SDKs read from it automatically.
Option 2: Start Evolution automatically in Claude Code
Claude Code Plugin (zero-code, recommended)
# In Claude Code console:
/plugin marketplace add Prismer-AI/PrismerCloud
/plugin install prismer@prismer-cloud
/reload-plugins
On first session, the plugin will guide you through setup:
/prismer:prismer-setup
# Auto opens browser for one-click sign-in (zero copy-paste, 30 seconds)
9 hooks run automatically — errors detected, strategies matched, outcomes recorded, web content cached, memory synced.
MCP Server (Claude Code / Cursor / Windsurf)
# Claude Code — one command
claude mcp add prismer -- npx -y @prismer/mcp-server
For Cursor / Windsurf, add to .cursor/mcp.json (or .windsurf/mcp.json):
{
"mcpServers": {
"prismer": {
"command": "npx",
"args": ["-y", "@prismer/mcp-server"],
"env": { "PRISMER_API_KEY": "sk-prismer-xxx" }
}
}
}
Reload the IDE — 29 tools available immediately (evolve_*, memory_*, context_*, skill_*).
No API key yet? Run
npx @prismer/sdk setupfirst.
Works Everywhere
<table> <tr><td><strong>Agent Integrations</strong></td><td><strong>Install</strong></td><td><strong>What it does</strong></td></tr> <tr><td>Claude Code Plugin</td><td><code>/plugin marketplace add Prismer-AI/PrismerCloud</code></td><td>9-hook auto-evolution, context cache, memory sync, skill sync</td></tr> <tr><td>MCP Server</td><td><code>npx -y @prismer/mcp-server</code></td><td>33 tools for Claude Code / Cursor / Windsurf</td></tr> <tr><td>OpenCode Plugin</td><td><code>opencode plugins install @prismer/opencode-plugin</code></td><td>Evolution hooks for OpenCode</td></tr> <tr><td>OpenClaw Channel</td><td><code>openclaw plugins install @prismer/openclaw-channel</code></td><td>IM channel + 14 agent tools</td></tr> </table> <table> <tr><td><strong>SDKs</strong></td><td><strong>Install</strong></td></tr> <tr><td>TypeScript / JavaScript</td><td><code>npm i @prismer/sdk</code></td></tr> <tr><td>Python</td><td><code>pip install prismer</code></td></tr> <tr><td>Go</td><td><code>go get github.com/Prismer-AI/PrismerCloud/sdk/prismer-cloud/golang</code></td></tr> <tr><td>Rust</td><td><code>cargo add prismer-sdk</code></td></tr> </table>Evolution Engine: How Agents Learn
The evolution layer uses Thompson Sampling with Hierarchical Bayesian priors to select the best strategy for any error signal. Each outcome feeds back into the model — the more agents use it, the smarter every recommendation becomes.

Agent A hits error:timeout → Prismer suggests "exponential backoff" (confidence: 0.85)
Agent A applies fix, succeeds → outcome recorded, gene score bumped
Agent B hits error:timeout → same fix, now confidence: 0.91
Network effect: every agent's success improves every other agent's accuracy
How it works:
- Signal detection — 13 error patterns classified from tool output (build failures, TypeScript errors, timeouts, etc.)
- Gene matching — Three-layer scoring: exact tag match → category prefix match → semantic similarity
- Thompson Sampling — Beta posterior sampling with hierarchical pooling (local agent data + global cross-agent prior)
- Outcome recording — Success/failure updates edge counts, quality-gated to prevent spam
- Person-Level Sync — All agent instances of the same user share genes (digital twin foundation)
Key properties:
- Sub-millisecond local — cached genes require no network
- 267ms propagation — one agent learns, all agents benef
