DriveMind
Security checks across static analysis, malware telemetry, and agentic risk
Overview
DriveMind is an instruction-only reliability and review framework with no code, credentials, or install steps; the only notable item is its encouragement to create reusable lessons or diary-style memory.
This appears safe to install as an instruction-only skill. Before using its review or diary features, check that any saved lessons are accurate, limited to the task, and do not preserve sensitive details unnecessarily.
Static analysis
No static analysis findings were reported for this release.
VirusTotal
VirusTotal findings are pending for this skill version.
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 the agent stores lessons, inaccurate or overly broad lessons could shape future responses.
This instructs the agent to create reusable lessons or continuity notes. That is purpose-aligned, but persistent summaries can affect later tasks if saved and reused.
Persist stable lessons, not raw noise. Use `templates/distill-template.md` for reusable lessons and `templates/diary-template.md` for daily continuity.
Review any saved lessons or next-time rules for accuracy, scope, and sensitive details before allowing them to persist.