Back to skill

Security audit

Facebook for Shopify via Agiexa

Security checks for vulnerabilities and agentic risk

Overview

This skill previews a future Facebook integration and offers an optional waitlist signup that requires your explicit permission.

Installing this skill does not enable Facebook actions. If you join the waitlist, Agiexa receives the details you confirm; omit the optional URL and message if you prefer. The separate future integration's security assurances are claims in the documentation, not guarantees established by this package.

Vulnerability Patterns
  • 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
  • Unauthorized Access and Privilege EscalationObtains permissions beyond the task's legitimate needs
Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (6)

Description-Behavior Mismatch

High
Category
MCP Tool Poisoning
Confidence
99% confidence
Finding

The supplied code performs only a waitlist submission. It does not publish posts, manage comments, retrieve insights, or implement the advertised Facebook tools. Although the endpoint belongs to the declared integration provider, waitlist enrollment is a materially different purpose rather than a supporting implementation detail of the advertised operations.

Content

No source excerpt is available for this finding.

Credential Access

High
Category
Privilege Escalation
Confidence
70% confidence
Finding

Code accesses credential files (SSH keys, AWS credentials, etc.). This could indicate credential theft attempts.

Content

Scanner excerpt · references/tools.md (reported line 21)May include surrounding context.

md
| `FACEBOOK_GET_COMMENT` | Retrieves details of a specific Facebook comment |
| `FACEBOOK_GET_COMMENTS` | Retrieves comments from a Facebook post or comment (for replies) |
| `FACEBOOK_GET_CONVERSATION_MESSAGES` | Retrieves messages from a specific conversation |
| `FACEBOOK_GET_CURRENT_USER` | Validates the access token and retrieves the authenticated user's own profile via /me |
| `FACEBOOK_GET_MESSAGE_DETAILS` | Retrieves details of a specific message sent or received by the Page |
| `FACEBOOK_GET_PAGE_CONVERSATIONS` | Retrieves a list of conversations between users and the Page |
| `FACEBOOK_GET_PAGE_DETAILS` | Fetches details about a specific Facebook Page |

Credential Access

High
Category
Privilege Escalation
Confidence
70% confidence
Finding

Code accesses credential files (SSH keys, AWS credentials, etc.). This could indicate credential theft attempts.

Content

Scanner excerpt · references/tools.md (reported line 35)May include surrounding context.

md
| `FACEBOOK_GET_POST_INSIGHTS` | Retrieves analytics and insights for a specific Facebook post |
| `FACEBOOK_GET_SCHEDULED_POSTS` | Retrieves scheduled and unpublished posts for a Facebook Page |
| `FACEBOOK_LIKE_POST_OR_COMMENT` | Adds a LIKE reaction to a Facebook post or comment |
| `FACEBOOK_LIST_MANAGED_PAGES` | Retrieves a list of Facebook Pages that the user manages (not personal profiles), including page details, access tokens, and tasks |
| `FACEBOOK_MARK_MESSAGE_SEEN` | Marks a user's message as seen by the Page, visibly updating the read status in the user's conversation |
| `FACEBOOK_PUBLISH_SCHEDULED_POST` | Publishes a previously scheduled or unpublished Facebook post immediately |
| `FACEBOOK_REMOVE_PAGE_TASK` | Removes a user's tasks/access from a specific Facebook Page |

External Transmission

Medium
Category
Data Exfiltration
Confidence
60% confidence
Finding

Data is being sent to an external URL. This could be legitimate telemetry or data exfiltration. Manual review is recommended.

Content

Scanner excerpt · SKILL.md (reported line 9)May include surrounding context.

md
openclaw:
    requires:
      bins:
        - curl
    homepage: https://agiexa.com/integrations/facebook
---

The manifest advertises Facebook operations, but the task workflow defers them and implements waitlist registration instead.

Medium
Category
Excessive Agency
Confidence
99% confidence
Finding

The description at L003 presents publishing posts, handling comments, and checking insights as available capabilities. However, L070 directs the agent to explain that these capabilities are not yet available, and the supplied executable workflow posts user contact details to a launch waitlist rather than performing Facebook operations. Explicit consent limits the data-submission risk but does not resolve the mismatch with the advertised purpose.

Content

No source excerpt is available for this finding.

The privacy assurance contradicts the waitlist payload's inclusion of store information.

Low
Category
Excessive Agency
Confidence
97% confidence
Finding

L114 states that nothing about the user's store or conversation is sent. Yet L085 requests a Shopify store URL, and L095–L098 include that URL and a user-provided message in the outbound payload. Although these fields are optional and submission requires consent, the categorical privacy assurance contradicts the documented request.

Content

No source excerpt is available for this finding.