Proactive Claw
Security checks across static analysis, malware telemetry, and agentic risk
Overview
This appears to be a purpose-aligned calendar productivity skill, but it needs broad calendar access, stores local learning data, and can run as a user-started proactive daemon.
Before installing, make sure you are comfortable granting broad calendar access and storing local productivity memory. Start with dry-run/preview flows, keep confirmation mode enabled, verify the Actions calendar configuration, and only run the background daemon if you want ongoing proactive behavior.
Static analysis
No static analysis findings were reported for this release.
VirusTotal
63/63 vendors flagged this skill as clean.
Risk analysis
Artifact-based informational review of SKILL.md, metadata, install specs, static scan signals, and capability signals. ClawScan does not execute the skill or run runtime probes.
If enabled, the skill can add, update, and clean up its own action-calendar events based on your schedule.
The skill can read calendars and create/update calendar entries, but the documented scope says writes are intended to be limited to the Proactive Claw Actions calendar.
Calendar read requests (event list, calendar list), write requests to Actions calendar only
Keep the default confirmation mode until comfortable, verify the Actions calendar ID in config.json, and use dry-run/audit commands before allowing routine changes.
Granting this permission lets the skill access calendar data and technically gives broad calendar authority, even though the artifacts say writes are code-restricted.
The skill uses delegated Google Calendar credentials and the full calendar OAuth scope, which is expected for this integration but grants broad account-level calendar authority.
`token.json` | Google OAuth access + refresh token ... Google OAuth Scope: `https://www.googleapis.com/auth/calendar`
Only install if you are comfortable granting calendar access; revoke access from your Google account permissions page or remove Nextcloud app credentials if you stop using it.
Local memory may contain sensitive work patterns, contacts, meeting outcomes, and preferences that influence later recommendations.
The skill keeps persistent local memory about meetings, contacts, rules, and user responses so it can personalize future suggestions.
`memory.db` | Meeting outcomes, rules, policies, contacts, notification logs, energy scores, proactivity scores ... default: 90 days
Review exported data periodically, tune retention settings, and delete the local state folder if you no longer want the skill to retain learned context.
While the daemon loop is running, it can continue scanning calendars and generating proactive reminders or actions.
The skill supports long-running proactive behavior, but the provided instructions describe it as manual or explicitly opt-in.
Background (manual): run local daemon yourself (`python3 scripts/daemon.py --loop`) ... Autonomous (advanced): explicit opt-in only
Do not run the loop or enable autonomous settings unless you want continuous behavior; keep `daemon_enabled` false and `max_autonomy_level` set to `confirm` for conservative use.