agent-cost-monitor

What This Skill Does

The agent-cost-monitor skill is a comprehensive financial observability tool for your OpenClaw ecosystem. By acting as a financial auditor for your AI operations, this skill tracks real-time token consumption and translates raw usage data into actionable financial metrics. It periodically polls your system's session_status, calculates costs based on current model pricing, and reconciles these figures against your defined budgetary constraints. Beyond passive reporting, the skill is designed to prevent 'bill shock' by providing proactive alerts when budget thresholds are approached and offering specific optimization strategies—such as recommending model switching or adjusting heartbeat intervals—to maximize ROI without sacrificing performance.

Installation

To add this capability to your agent, execute the following command in your terminal or via the ClawHub CLI:

clawhub install openclaw/skills/skills/neal-collab/agent-cost-monitor

Once installed, ensure your agent has read access to the session logs and environment variables where your budget thresholds are defined. For automated reporting, integrate the provided cron job schema into your OpenClaw orchestration configuration to receive daily financial summaries directly in your preferred notification channel.

Use Cases

• Proactive Budgeting: Set weekly expenditure limits and receive automated warnings before you exceed them.
• Model Optimization: Identify which tasks can be downgraded from premium models (e.g., Opus) to efficient alternatives (e.g., Haiku or local Ollama instances) without impacting task success rates.
• Usage Auditing: Determine which specific agents are responsible for the highest token consumption within a team or project group.
• Cost Recovery Analysis: Calculate the 'cost-per-task' for your automated workflows to determine the economic viability of specific agent-driven processes.

Example Prompts

1. "Run a full audit on all active agents and show me the top three contributors to my monthly bill."
2. "If I keep running the current workload with these settings, will I hit my $50 weekly budget limit?"
3. "Show me which agents are currently using expensive models for tasks that could be handled by local Ollama models."

Tips & Limitations

• Pricing Updates: The skill uses a reference lookup table; ensure your local version is updated if model pricing structures change.
• Ollama/Local Tracking: While Ollama is free, tracking still consumes overhead; ensure the monitor is configured to exclude local resources from monetary calculation to keep reports clean.
• Accuracy: Cost estimations are based on standardized API usage rates and may vary slightly based on specific provider billing cycles or account-level discounts.