Back to skill
Skillv1.0.0

ClawScan security

Memory Distill · ClawHub's context-aware review of the artifact, metadata, and declared behavior.

Scanner verdict

BenignMar 20, 2026, 3:12 PM
Verdict
benign
Confidence
high
Model
gpt-5-mini
Summary
The skill is an instruction-only 'memory distill' helper whose required actions (scanning conversation history, extracting summaries, and writing local memory files or scheduling cron/heartbeat triggers) match its stated purpose and request no unrelated credentials or installs.
Guidance
This skill appears coherent: it will read your conversation history and write summarized memory files and can schedule itself via Cron or HEARTBEAT. Before enabling it, review where memory files (MEMORY.md and memory/YYYY-MM-DD.md) will be stored, ensure they don't inadvertently contain secrets, set retentionDays and keep autoClean/autoReset disabled unless you want automated deletion/reset, and only add cron/heartbeat entries if you trust the agent to run these tasks. If you need stronger privacy, ask for automatic redaction/encryption or avoid persisting sensitive conversations.

Review Dimensions

Purpose & Capability
okThe name/description (distilling conversation memory into structured files, scheduling via Cron/HEARTBEAT) align with the instructions: scanning session history, extracting decisions/tasks/knowledge, writing MEMORY.md and memory/YYYY-MM-DD.md, and configuring cron or heartbeat triggers. No unrelated binaries, env vars, or external services are requested.
Instruction Scope
noteInstructions explicitly direct the agent to scan the current conversation history and write structured memory files and reports — this is coherent with the purpose. Note: the skill will read user conversation content (potentially sensitive) and persist summaries to workspace files; SKILL.md leaves sensitive-data handling largely to manual marking and does not provide automated redaction or encryption guidance.
Install Mechanism
okInstruction-only skill with no install spec or external downloads. Lowest-risk delivery method; nothing is written to disk by an installer beyond the agent following file-write instructions at runtime.
Credentials
okNo environment variables, credentials, or config paths are required. The requested scope (local file writes, cron/heartbeat integration) is proportionate to the described functionality.
Persistence & Privilege
okalways:false and normal agent invocation are used. The skill does not request permanent platform-wide privileges or modify other skills' configs. It suggests creating cron jobs and updating HEARTBEAT.md, which is expected for scheduled automation but should be done intentionally by the user.