Context Manager

• Risk level: Low
• Analyzed: Apr 17, 2026, 1:30 AM
• Findings: 0
• Version: 1.1.0

Overview

The skill's files and runtime instructions are internally consistent with a personal context/knowledge-management tool; it operates on a local data/ directory, includes an optional entropy-reduce script that can mark/delete low-value notes, and does not request external credentials or include suspicious network endpoints.

This skill appears coherent and operates on its own data/ directory. Important things to check before installing or allowing autonomous runs: 1) Inspect scripts/entropy_reduce.py and confirm the data path it uses (it defaults to a data/ folder relative to the skill). 2) When running entropy reduction, run with --dry-run first to review the generated report; only use --force after you confirm deletions. 3) If you plan to enable integrations (WeChat, 飞书, webhooks), provide credentials only after verifying the integration code and understanding where/when data is transmitted. 4) If you do not want the agent to run scripts autonomously, restrict the skill's Exec permission or require explicit user confirmation before running the entropy-reduce action.