Monitor local agent sessions and costs with ClawMetry
Use ClawMetry when an operator needs a local dashboard for agent sessions, token spend, tool calls, alerts, and governance across multiple runtimes.
npx skills add agentskillexchange/skills --skill monitor-local-agent-sessions-and-costs-with-clawmetry
Use ClawMetry when an operator needs to observe and govern local AI agent work across several runtimes from one dashboard. The tool auto-detects supported agents, reads local session and log data, shows transcripts, token spend, tool calls, reasoning and event streams, health signals, alerts, approvals, and OpenTelemetry export paths. Invoke this when supervising active coding agents, investigating cost spikes, checking stalled sessions, auditing tool usage, or setting pre-execution guardrails across Claude Code, Codex, Cursor, Gemini CLI, OpenClaw, Goose, Aider, and related runtimes. The boundary is observability and governance for agent sessions; it is not a generic APM platform, generic analytics dashboard, or SDK listing.
What this skill actually does
Inputs and prerequisites: Python 3.8+, clawmetry package, at least one supported local agent runtime, optional license or self-hosted configuration for paid runtime coverage.
Setup notes: Install with pip install clawmetry and run clawmetry. The dashboard opens locally at http://localhost:8900 and auto-detects supported local agent runtimes. Docker and one-line shell installer options are documented upstream.
Source and verification boundary: use https://clawmetry.com as the canonical reference before running the workflow; keep commands, API calls, CLI usage, and generated outputs reviewable against that upstream source.
Framework fit: publish this as a Multi-Framework workflow only when the operator can invoke the documented toolchain directly, rather than treating the upstream project as a generic product listing.