Watch local agent token usage and session traces with TokenTelemetry
Use TokenTelemetry to read local Claude Code, Codex, Gemini CLI, Hermes, and other agent logs into a dashboard for token usage, costs, tool calls, project activity, budgets, and session traces.
npx skills add agentskillexchange/skills --skill watch-local-agent-token-usage-and-session-traces-with-tokentelemetry
Use TokenTelemetry when an operator runs multiple coding or autonomous agents and needs a local dashboard for token burn, session traces, tool calls, model efficiency, project activity, budget alerts, and plan-limit visibility. TokenTelemetry reads the log files agents already write, binds locally by default, and supports Codex, Claude Code, Gemini CLI, Cursor, Hermes Agent, and other agent surfaces without requiring application instrumentation.
What this skill actually does
The repeatable workflow is concrete: install TokenTelemetry, run it on the same machine or host where agent logs live, let it auto-detect supported agents, review token and cost dashboards, inspect waterfall traces, compare projects or models, and configure observational budgets or remote access only when needed. Invoke it instead of normal provider dashboards when the question is cross-agent, project-local, or trace-level usage analysis from files already on the machine. The scope boundary is local observability for existing agent logs, not blocking agent execution, proxying model calls, or replacing full application observability platforms.