openclaw-audit-watchdog

Automated daily security audits for OpenClaw agents with DM delivery and optional email reporting. Runs deep audits, creates or updates a recurring cron job,...

This package is internally consistent with its stated purpose, but check the following before enabling: 1) Verify you trust the openclaw CLI and the skill source (README suggests downloading a release — verify signatures/URL if you use that path). 2) Review and confirm DM and optional email recipients (PROMPTSEC_DM_CHANNEL, PROMPTSEC_DM_TO, PROMPTSEC_EMAIL_TO) to avoid sending audit output to unintended parties. 3) Inspect which environment variables will be persisted into the cron payload (setup_cron.mjs lists PERSISTED_ENV_KEYS) and avoid baking any secrets you don't want stored in a recurring job. 4) If you enable suppressions, understand that they require both the --enable-suppressions flag and a config file sentinel; review any suppression files before enabling. 5) Optionally inspect the scripts (runner.sh, setup_cron.mjs, send_smtp.mjs) locally to confirm behavior and delivery paths (send_smtp can be pointed to a remote relay if you set PROMPTSEC_SMTP_HOST). If any of these checks fail or you cannot verify the release origin, do not enable the unattended cron.

Type: OpenClaw Skill Name: openclaw-audit-watchdog Version: 0.1.4 The openclaw-audit-watchdog skill is a security monitoring tool designed to automate daily audits and report findings via DM or email. The codebase demonstrates high security maturity, featuring robust input sanitization for shell commands (setup_cron.mjs), email header cleaning to prevent injection (sendmail_report.sh, send_smtp.mjs), and a defense-in-depth suppression mechanism that requires both a CLI flag and a configuration sentinel to activate. All data delivery destinations are user-configured, and the instructions in SKILL.md are transparently aligned with the stated purpose of scheduling and executing security audits.