Back to skill

Security audit

NotFair Meta Ads Agent

Security checks for vulnerabilities and agentic risk

Overview

This skill is a disclosed Meta Ads management integration, but users should treat it as sensitive because it installs and connects an external plugin with ad-account write capabilities.

Install only if you trust NotFair and are comfortable connecting it to Meta Ads. Use least-privilege Meta permissions where possible, review proposed writes carefully, and require explicit approval for budget, targeting, creation, enablement, or creative changes.

Vulnerability Patterns
  • Insecure DependenciesIntroduces malicious components through unsafe dependency sources
  • 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 (1)

T08 · Insecure Dependencies

Warning
Location
SKILL.md:34
Finding

Unpinned Third-Party Plugin Installation

Content
View full analysis

Vulnerability Details

File Location: SKILL.md, lines 34–37 and 48–50
Vulnerability Type: Unpinned and unauditable third-party dependency
Risk Level: Medium

Vulnerable Code

yaml
install:
  - id: openclaw-notfair
    kind: node
    label: "NotFair Meta Ads Plugin"
text
openclaw plugins install clawhub:openclaw-notfair
openclaw plugins enable openclaw-notfair
openclaw config set 'plugins.entries.openclaw-notfair.config.mcpUrl' '"https://notfair.co/api/mcp/meta_ads"' --strict-json

Technical Analysis

The documented installation command references clawhub:openclaw-notfair without an immutable version, commit, integrity hash, or signed artifact. The plugin implementation is not included in the audited project, so its executable behavior and enforcement of the documented approval model cannot be verified from SKILL.md.

This creates a supply-chain risk: the effective code installed by users may change after this skill has been reviewed. Compromise of the package registry entry, publisher account, distribution infrastructure, or a future release could cause users to install code different from the reviewed instructions.

The configured hosted MCP endpoint further places authenticated advertising operations behind components that are not available in this project for inspection. This is disclosed and consistent with the skill's purpose, but it increases the consequences of dependency compromise.

Attack Path

  1. An attacker compromises the plugin publisher account, package entry, or distribution infrastructure, or publishes a malicious mutable update.
  2. A user follows the documented unversioned installation command.
  3. OpenClaw resolves and installs the attacker-controlled plugin version.
  4. The user enables the plugin and authenticates it for the NotFair and Meta Ads workflow.
  5. The malicious plugin abuses its execution context or authenticated advertising access to co ...[truncated 853 chars]
Remediation
View remediation

Remediation Suggestions

  1. Pin the plugin to an immutable, reviewed version rather than installing an unspecified current release.
  2. Publish and verify a cryptographic integrity hash or signed provenance for the exact artifact.
  3. Include the plugin source in the review scope or link the installation to a reproducible, immutable source commit.
  4. Document the plugin's requested host permissions, credential access, data transmission, and MCP authorization boundaries.
  5. Grant the plugin and associated tokens only the minimum Meta Ads permissions required for the requested task.
  6. Enforce approval for money-affecting writes in a trusted local authorization layer rather than relying solely on the remote plugin or MCP service.
  7. Separate read-only and write-capable credentials where supported, and keep write access disabled until explicitly needed.
  8. Monitor installed artifact integrity and require review before accepting dependency updates.
Vulnerability Patterns
  • 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
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep