Parse state.db and agent.log to audit local token usage and costs.
npx skills add https://github.com/qiuhaomem/HawkEye-Mem --skill hermes-token-audit
Please help me install this Agent Skill. Skill: hermes-token-audit Source: https://github.com/qiuhaomem/HawkEye-Mem/tree/main/skills/hermes-token-audit Command: npx skills add https://github.com/qiuhaomem/HawkEye-Mem --skill hermes-token-audit
Token audit solves the lack of local visibility into token consumption and costs by performing offline, privacy-preserving analysis on HawkEye-Mem data.
Run hawk-eye-mem --token-audit to perform a local token audit.
High-intent search queries and answers about installing and using this skill.
To audit local token usage, you can run an offline analysis on state.db and agent.log files. This process parses local data to compute totals, source distribution, and waste patterns without sending any data externally.
On-device token auditing parses local memory state files to compute token costs and usage trends locally. It operates entirely offline using Python standard library modules, ensuring data never leaves your machine for privacy-preserving governance.
Yes, you can analyze token waste and costs using only the Python standard library. The audit relies solely on built-in modules like sqlite3, json, datetime, re, and csv, requiring zero external dependencies to parse state data.
To get structured JSON output for token cost trend analysis, run a local token audit on your state data. The audit computes distribution and waste across a defined time window, outputting structured JSON formatted for AI agents.
No, the token audit requires zero external dependencies and never sends data externally. It uses only the Python standard library to perform on-device analysis, maintaining strict privacy and operating offline.
jonnytacos8Convert text to speech locally using sherpa-onnx across darwin, linux, and win32.
filipemottaAnalyze cloud spend to identify waste and recommend cost optimizations.
baphledEstimate and track token costs for AI-assisted work sessions.
vinwangGenerate on-device text and structured outputs with Apple FoundationModels for iOS apps.
