Where Am I Burning Tokens?

Reads only ~/.openclaw/agents/main/sessions/sessions.json to show exactly where OpenClaw tokens and estimated cost are going by session type. Read-only diagn...

This skill is internally consistent with its stated purpose: it reads ~/.openclaw/agents/main/sessions/sessions.json and summarizes tokens/costs. Two practical precautions before installing or invoking it: (1) be aware the skill advertises 'read-only' but offers an optional write (it will create/append ~/.openclaw/workspace/memory/token-diet-log.md) if you explicitly ask to track improvements — don't request that unless you want a file written. (2) As with any instruction-only skill, its runtime behavior depends on the agent implementation; ensure your agent runtime enforces file/network access policies you trust (the skill claims it won't exfiltrate data, but the platform executing the instructions controls actual IO/network). If you want extra assurance, inspect the actual sessions.json contents to confirm it does not contain secrets you wouldn't want read, and only enable the skill when comfortable with it having read access to that single path.

Type: OpenClaw Skill Name: whereamiburningtokens Version: 1.1.0 The skill is a diagnostic tool designed to analyze token usage and costs by reading the local `sessions.json` file. It operates within a strictly defined safety boundary, performing read-only operations on session data and optional logging of improvements in the workspace directory. No evidence of data exfiltration, unauthorized execution, or malicious intent was found in SKILL.md or the associated metadata.