记忆秘书

提供智能记忆管理、重复任务检测、成功案例提取、工作模式分析和智能提醒生成的助手工具。

This package looks like a local memory-analysis tool, but there are several red flags you should resolve before installing or running it: - Missing files: src/__init__.py imports modules (smart_adaptive, daily_check, pilot_check) that are not present in the manifest, and tests reference many scripts and a memory/ directory that are not packaged. Ask the author for the full package or a corrected manifest. - Read vs write mismatch: SKILL.md says it won't modify memory files, yet the code creates a secretary data directory and writes JSON/SQLite files inside memory/. If you want read-only analysis, insist the author document exactly what files are created and where. - Workspace paths: the default workspace_root is /home/admin/openclaw/workspace; explicitly pass a workspace_root you control when constructing MemorySecretaryLite to avoid accidental reads/writes outside your intended area. - Safety steps before running: (1) Inspect all src/*.py for network calls or subprocess.exec usage (none were obvious in the truncated code, but check the full files). (2) Run the package in an isolated environment (container or VM) and with a backup of your real memory/ directory. (3) Prefer running with a temporary test workspace that contains representative memory/ files. (4) Run the provided tests locally and review failures — many tests expect files that are missing, which could indicate incomplete packaging or hidden runtime assumptions. If the author can supply the missing modules and explain the on-disk writes (and update SKILL.md to accurately reflect them), the package would be more coherent; until then treat it as incomplete and proceed with caution.

Type: OpenClaw Skill Name: memory-secretatry Version: 1.0.0 The 'Memory Secretary' skill is a legitimate utility designed to manage and optimize agent memory files within the OpenClaw workspace. The core logic in 'src/memory_secretary_lite.py' uses standard Python libraries to perform file integrity checks, similarity analysis via difflib, and pattern recognition using regular expressions. No evidence of data exfiltration, unauthorized network access, or malicious code execution was found. While the bundle references some missing modules (e.g., smart_adaptive.py) and contains a minor typo in the metadata slug, the provided code is functional, well-documented, and strictly follows its stated purpose of local memory management.