Session Token Ledger

Analyze local OpenClaw session token usage from a generated SQLite ledger and markdown summaries. Use when the user asks for a token audit, token体检报告, contex...

This skill appears internally consistent and does what it says: it scans local OpenClaw session transcript files and creates a local SQLite ledger and markdown reports. Before installing or enabling the optional hook: (1) review the included scripts (rebuild_sqlite.py and report.py) yourself — they will read all transcript files under agents/.../sessions and write files into the skill's assets directory; session transcripts can contain sensitive messages or secrets; (2) run the scripts manually first (python3 scripts/rebuild_sqlite.py and python3 scripts/report.py) to inspect outputs before enabling automatic hooks; (3) only enable the hook if you want automatic rebuilds on /new, /reset, /stop, or gateway startup; (4) consider filesystem permissions and where the skill writes outputs (assets/ and .generated/ report folder); and (5) if you run in a multi-user environment, be cautious about sharing the generated artifacts. No network exfiltration or credential requests were found, so the primary risk is accidental local exposure of transcript content — protect the generated files accordingly.

Type: OpenClaw Skill Name: session-token-ledger Version: 1.0.0 The session-token-ledger skill is a local utility designed to audit and analyze LLM token usage from OpenClaw session transcripts. It parses local JSONL log files, populates a SQLite database (session_tokens.db), and generates markdown reports via Python scripts (rebuild_sqlite.py and report.py). The code operates entirely on local files within the OpenClaw directory structure, contains no network exfiltration logic, and uses standard platform features for automation hooks without any evidence of malicious intent or prompt-injection attacks.