memddc

面向团队协作的项目文档管理与代码迭代工具，支持自动文档生成、DDD模型管理、记忆压缩和智能变更同步。

This skill appears to do what it claims (scan the repo, build an index, generate docs, and update code/docs based on a mem-snapshot), but it performs broad repository scans and reads IDE indexes, config files and VCS logs — which can include secrets or sensitive info. Before installing/use: 1) Review and vet the README/SKILL.md locally; run it on a non-production or sample repository first. 2) Add strict excludes to .memddc/config.json (or to your repository) to prevent scanning of .env, credential files, keystores, CI secrets, and other sensitive paths. 3) Ensure .memddc/ is included in .gitignore so snapshots and logs aren't accidentally committed. 4) Prefer manual invocation (avoid scheduling/autonomous runs) until you verify what artifacts are produced. 5) Because the publisher/source is unknown (no homepage), inspect the example outputs under example/.memddc and verify there are no hidden remote endpoints or implicit uploads. If you need help auditing the exact set of files the skill will read or the contents of a mem-snapshot.json produced in your repo, run initialization in an isolated clone and review the snapshot before allowing any external transmission.

Type: OpenClaw Skill Name: memddc-ai-skill Version: 1.0.2 MemDDC is a project documentation and code iteration tool designed to optimize AI agent performance by maintaining a structured 'memory snapshot' (mem-snapshot.json). It automates the generation of DDD models, architecture diagrams, and API documentation by scanning the codebase and analyzing VCS (Git/SVN) logs. While the skill requires broad file system access and executes Git commands (e.g., `git log`, `git diff`) to track changes and build context, these actions are strictly aligned with its stated purpose of project governance and token-usage reduction. No evidence of data exfiltration, malicious execution, or harmful prompt injection was found in the SKILL.md or README.md files.