Clarence

Clarence is a production-proven AI Operator persona built from the ground up to run your business autonomously. He operates as a full-stack business operator: qualifying leads, managing outbound email, monitoring revenue, auditing payments, handling client communications, and maintaining a three-layer persistent memory system that actually works at scale.

Clarence has managed live Stripe revenue, run outbound email sequences, scored and qualified leads, and operated 24/7 on a cron schedule — all autonomously with appropriate human oversight.

🎁 All Skills Included The Clarence persona includes every skill in the Clarence Protocol — no separate purchases needed:

• 3-Layer Memory System — working memory + episodic log + long-term knowledge base with nightly consolidation
• Gmail Operator — autonomous email management with OAuth2 and daily send scheduling
• Stripe Monitor — revenue tracking, subscription auditing, and webhook failure recovery
• Heartbeat System — integration health checks every 5 minutes with auto-escalation
• Lead Engine — automated lead scoring, qualification, and outbound sequencing

Proven at scale:

• Escalation rules calibrated to interrupt only when stakes are real — financial thresholds, data risks, consecutive failures
• Three-layer memory with nightly consolidation — working context resets daily, but nothing is ever lost
• Heartbeat system that detects integration failures and escalates immediately without being asked
• Cron discipline that prevents the #1 autonomous AI failure mode: silent task failures with no recovery

What makes Clarence different from a generic prompt:

• Production-proven patterns — not theory, but operational knowledge distilled from real deployment and real failure modes
• Anti-pattern awareness — knows what NOT to do: don't post to social media without a review queue, don't execute financial writes without a BILLING_WRITE flag, don't guess on high-stakes decisions, don't deploy SOUL.md changes without running the validator
• 3-layer memory — working memory + episodic log + semantic knowledge base, with automatic nightly consolidation, pattern extraction, and memory decay so context stays fresh and lean
• Escalation discipline — calibrated thresholds for when to stop and ask vs. execute autonomously, version-controlled in SOUL.md like production code
• Autonomous operations — heartbeat monitoring, Stripe auditing, lead scoring, email sequencing, and nightly memory lifecycle — all running without being prompted
• Full tool configuration — pre-configured Gmail OAuth2, Stripe read/write gates, terminal execution with allowlists, and n8n workflow integration out of the box

Includes pre-configured cron schedules (5-min heartbeat, 2-hour lead scoring, daily memory consolidation, weekly knowledge synthesis), specialized skill documents (memory architecture, lead operations), and a complete README with installation guide. Designed for OpenClaw. Runs on Claude (Opus, Sonnet, or any supported model), Gemini, Kimi-K2.5 & others. The persona patterns are model-agnostic — the value is in the operational knowledge.