Skill flagged — suspicious patterns detected
ClawHub Security flagged this skill as suspicious. Review the scan results before using.
Meeting Assistant
v1.0.0从会议录音或文字生成纪要,提取关键决策和待办事项,支持会议安排和提醒通知管理。
⭐ 0· 27·0 current·0 all-time
MIT-0
Download zip
LicenseMIT-0 · Free to use, modify, and redistribute. No attribution required.
Security Scan
OpenClaw
Suspicious
medium confidencePurpose & Capability
Name/description match the code's functions (minutes generation, action-item extraction, templates, simple scheduling queries). The skill does not request credentials or binaries, which is appropriate for its stated offline functionality. However, SKILL.md documents storing data under ~/.openclaw/workspace/meeting-assistant/meetings.json, while the code persists data relative to the skill package (Path(__file__).parent). This mismatch could surprise users about where their data is saved.
Instruction Scope
SKILL.md lists file read/write and timed tasks (reminders) and mentions calendar integration as a planned feature; the provided code reads/writes local JSON and implements parsing, templates, listing and basic scheduling queries, but I found no implementation of external calendar API calls or reminder delivery in the visible code. The SKILL.md and code disagree about the storage path and about reminder/calendar behavior, which is scope creep or incomplete implementation rather than malicious behavior.
Install Mechanism
There is no install specification (instruction-only at registry level) and no network-based installer. The skill includes a Python file and local JSON; nothing is downloaded or extracted at install time — low install risk.
Credentials
The skill declares no required environment variables or credentials and the code does not reference external secrets in the visible portions. It only reads/writes JSON files in its directory. This is proportionate to an offline meeting assistant. Note: TASKS_FILE is declared but tasks.json is not present in the bundle and would be created in the same directory.
Persistence & Privilege
The skill does not request always:true and does not appear to modify other skills or system-wide settings. It will persist data locally (within the skill directory by code) but does not request elevated privileges.
What to consider before installing
What to consider before installing:
- Storage path mismatch: SKILL.md says data will be stored under ~/.openclaw/workspace/meeting-assistant/meetings.json, but the code writes to the skill's directory (Path(__file__).parent). Ask the author to confirm the intended storage location and prefer a user-controlled workspace path (or an explicit prompt) so personal data isn't kept in an unexpected place.
- Reminders / calendar integration: SKILL.md mentions reminders/timed tasks and calendar integration, but the visible code doesn't implement external calendar APIs or notification delivery. If you need calendar sync or notifications, request details about which APIs/endpoints will be used and what credentials (if any) are required.
- Data persistence: The code will create or update local JSON files (meetings.json and tasks.json) in the package directory. If you install in a shared or system-controlled location, data may be stored there. Consider running the skill in a sandbox or checking file writes first.
- Verify full source: The repository file was truncated in the provided bundle; review the remaining code to ensure there are no network calls, hidden endpoints, or subprocess calls. Ask for a complete listing or run a static check locally.
- Author and provenance: The skill has no homepage and the source is unknown. That reduces trust; prefer skills with identifiable authors or official repos. If you proceed, test with non-sensitive dummy data first.
- Security posture: Because no external credentials or network access are declared, immediate exfiltration risk appears low from the visible code, but the storage-path inconsistency and incomplete feature implementation are red flags for sloppy or unfinished code. If you need calendar/notification features, require explicit permission prompts and clear documentation of endpoints before granting credentials.Like a lobster shell, security has layers — review code before you run it.
latestvk974yq2rcekr3813vmd6ne7ywd841gst
License
MIT-0
Free to use, modify, and redistribute. No attribution required.
