notion-agent-memory

Structured memory system for AI agents using Notion. Use when setting up agent memory, discussing memory persistence, or helping agents remember context acro...

This skill appears to do what it says (templates + Notion integration) but contains risky recommendations around credential handling and background checks. Before installing or using it: 1) Do NOT store sensitive credentials in workspace docs like MEMORY.md or checked-in files. 2) If you use Notion integration, create a dedicated Notion integration with the minimum required permissions and store its token in a protected config file with restrictive file permissions (chmod 600), or use a secrets manager rather than a plain text file. 3) Review and remove/disable any instructions that would make the agent check email/calendar/send messages or run cron jobs unless you explicitly want that and have secured credentials. 4) Avoid putting tokens into version control or shared folders. 5) If you enable agent autonomy to perform external actions, restrict those capabilities and audit what the agent sends. If you want, I can point out exact lines to change to remove the credential-storage and background-check recommendations.

Type: OpenClaw Skill Name: notion-agent-memory Version: 1.0.1 The skill is classified as suspicious due to several high-risk capabilities and potential vulnerabilities. Specifically, the `SKILL.md`, `assets/MEMORY-TEMPLATE.md`, `references/act-framework.md`, and `references/notion-integration.md` files instruct the agent to store a Notion API key in `~/.config/notion/api_key` and then use shell commands like `$(cat ~/.config/notion/api_key)` or `$(grep ... | cut ...)` within `curl` calls. This pattern is vulnerable to shell injection if the content of the API key file can be controlled by an attacker. Additionally, the `assets/HEARTBEAT-TEMPLATE.md` file contains an instruction for the agent to 'Use cron jobs for background tasks', which, if interpreted as a direct command, could lead to unauthorized persistence on the system. While these actions are presented within the context of the skill's stated purpose (Notion integration and agent memory management), they represent significant security risks and vulnerabilities rather than clear malicious intent.