Clawdscan

What to check before installing: - Review the clawdscan.py source yourself (it's included). The tool reads session JSONL files under ~/.clawdbot and related dirs and can delete/archive sessions when run with --execute. - Default behavior is a preview (non-destructive), but the heartbeat/cron examples and skill.json include an auto-run/heartbeat config. Confirm how your platform handles the skill.json heartbeat block (does installing automatically schedule periodic runs?). - Do NOT enable automatic cleanup (CLAWDSCAN_AUTO_CLEANUP=true) until you have run several scans and validated the deletion behavior in a safe environment. - If you plan to run it on production data, take backups of your sessions first. Test cleanup in a staging environment and inspect the 'would remove' preview output carefully. - Expect the tool to write history/metadata files to your home (~/.clawdbot or ~/.clawdscanrc) if you enable history saving; if you prefer no persistence, run scans without --save-history or change paths in the config. - If you are concerned about network notifications, check any webhook/curl examples — they are optional and not required, but review any automation that posts scan results externally. Summary recommendation: the skill is broadly coherent with its stated purpose, but because it requests periodic execution and can perform destructive cleanup, treat it cautiously: audit the code, run in dry-run mode, and verify heartbeat/auto-cleanup settings before enabling automated runs.