notion-agent-memory

AdvisoryAudited by Static analysis on Apr 30, 2026.

Overview

No suspicious patterns detected.

Findings (0)

Artifact-based informational review of SKILL.md, metadata, install specs, static scan signals, and capability signals. ClawScan does not execute the skill or run runtime probes.

What this means

Raw secrets or instructions for using them could persist across sessions and be exposed to future agents, workspace collaborators, or anyone with access to the memory files or Notion pages.

Why it was flagged

The memory guidance explicitly encourages saving credentials and usage details in persistent notes, but does not specify redaction, secret-manager use, file permissions, or Notion sharing limits.

Skill content
Don't just save credentials → document how to use them with examples
Recommendation

Do not store raw credentials in MEMORY.md, AGENTS.md, or Notion. Store only references to a password manager, vault, or environment variable, and restrict the Notion integration to the minimum pages needed.

What this means

If implemented, an agent could keep monitoring messages, calendars, or notifications after the user’s immediate task is over.

Why it was flagged

The heartbeat template encourages recurring autonomous checks and background jobs that continue beyond the immediate session, without defining exact accounts, permissions, approval steps, or stop conditions.

Skill content
Check email/calendar every 2 hours while awake
- Monitor for urgent notifications
- Use cron jobs for background tasks that survive session breaks
Recommendation

Treat heartbeat/cron behavior as opt-in only. Define exact services, schedules, allowed actions, logging, and a clear way to disable it before enabling any background checks.

What this means

Future agent sessions may rely on saved instructions or context that the user did not intend to make permanent.

Why it was flagged

Persistent memory and self-updating instruction files are central to the skill, but they can influence future sessions if inaccurate, stale, overly sensitive, or malicious content is added.

Skill content
Every time you begin: 1. Read MEMORY.md ... Update the Instructions ... update this file. Next time starts smarter.
Recommendation

Review memory and instruction changes regularly, label sources and dates, avoid storing secrets or sensitive personal data, and treat saved memory as lower priority than direct user and system instructions.

What this means

An agent with the Notion token could add or modify memory entries in shared Notion pages or databases that the integration can access.

Why it was flagged

The Notion integration examples include creating and updating Notion pages. This is expected for a memory system, but it is still mutation authority over a third-party workspace.

Skill content
Create a Page ... curl -s -X POST "https://api.notion.com/v1/pages" ... Update a Page ... curl -s -X PATCH "https://api.notion.com/v1/pages/$PAGE_ID"
Recommendation

Share only the specific Notion pages/databases needed with the integration, review write actions, and use a dedicated low-privilege Notion integration.