Back to skill

Security audit

Second Brain Setup

Security checks for vulnerabilities and agentic risk

Overview

This skill is a narrowly scoped Notion setup helper, but it can create or move workspace objects without asking the user to confirm the exact location or existing databases first.

Before installing, make sure the Notion integration is shared only with the page where you want this setup to live. Avoid running it in a workspace where unrelated databases named Reading List or Inbox are visible to the integration, because the skill may move them automatically.

Vulnerability Patterns
  • Unauthorized Access and Privilege EscalationObtains permissions beyond the task's legitimate needs
  • Skill Instruction HijackingAlters the agent's session goals or safety constraints when the skill loads
  • Agent Memory PoisoningWrites attacker-controlled rules into memory that affect later sessions
  • Remote Payload Retrieval and ExecutionFetches external code whose behavior can change after review
  • Embedded Malicious CodeShips malicious scripts inside the skill and executes them locally
Findings (2)

T05 · Unauthorized Access and Privilege Escalation

Warning
Location
SKILL.md:94
Finding

Unconfirmed Relocation of Existing Notion Databases

Content
View full analysis
Remediation
View remediation

T05 · Unauthorized Access and Privilege Escalation

Note
Location
SKILL.md:57
Finding

Automatic Creation Under an Arbitrary Integration-Visible Page

Content
View full analysis
"}, "properties": {"title": "Second Brain"} } ] } ``` ### Technical Analysis The skill enumerates every page visible to the integration and heuristically selects any visible page as a write target. A page being visible to an integration does not establish that it is an appropriate location for new content. The preference for names such as `Home`, `Workspace`, or `IronClaw` is not an authorization check. In a workspace with multiple accessible areas, the selected page may belong to an unrelated team, shared project, or sensitive hierarchy. The skill does not present candidate parents to the user or require confirmation before creating the `Second Brain` page and its child databases. ### Attack Path 1. The Notion integration has access to multiple pages. 2. No existing page titled exactly `Second Brain` is found. 3. An unrelated or attacker-influenced visible page has a plausible title such as `Home` or `Workspace`. 4. The skill chooses that page as `INTEGRATION_PARENT_ID` using its heuristic. 5. The skill creates `Second Brain`, `Reading List`, and `Inbox` beneath that page. 6. Users with access to the selected parent may receive unintende ...[truncated 638 chars]
Remediation
View remediation
Vulnerability Patterns
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (2)

Unrestricted Tool Access

Medium
Category
Excessive Agency
Confidence
80% confidence
Finding

Skill grants unrestricted tool access without appropriate constraints. An agent with unfettered tool access can perform arbitrary actions including file modification, network requests, and code execution.

Content

Scanner excerpt · SKILL.md (reported line 33)May include surrounding context.

md
## STRICT RULES — read first

You will execute a **fixed sequence of Notion MCP tool calls**. You are NOT to invent pages, write markdown content, or call any tool not listed in the sequence below.

✅ **The only output of this skill:**
1. One **page** titled exactly `Second Brain` (no body content, no icon, no cover)

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
94% confidence
Finding

The skill can create and move existing Notion databases based on broad search results, but it does not require an explicit user confirmation before modifying workspace structure. In particular, the adopt-and-move behavior for any exact-match Reading List or Inbox database may relocate pre-existing content the user did not intend this setup to touch.

Content

No source excerpt is available for this finding.