ClawHub
Back to skill

Security audit

Twinfold

Security checks for vulnerabilities and agentic risk

Overview

This skill is mostly coherent for managing Twinfold, but it can publish social media content to live accounts and includes auto-publish/autopilot examples without clear confirmation gates.

Install only if you intend to let the agent manage Twinfold. Treat it as a Review item: require draft-first operation and explicit confirmation before publish, schedule, approve, delete, runAutopilot, or any autoPublish:true action, especially across multiple connected accounts or languages.

A.I.G

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

SkillSpector

Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Findings (7)

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
97% confidence
Finding

The trigger list is very broad and includes generic terms like "post," "content," and platform names, which can cause the skill to activate in contexts where the user did not specifically intend to interact with Twinfold. Because this skill can create, schedule, and publish external social media content, unintended invocation increases the chance of accidental outbound actions against the user's connected accounts.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
93% confidence
Finding

The skill description advertises workflows for posting, scheduling, autopilot, and account interaction without clearly warning that these operations can cause real external side effects on connected social platforms. In a tool that controls publication pipelines, omission of explicit confirmation guidance makes accidental posting or mass content generation materially more likely.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
92% confidence
Finding

The autopilot section describes runAutopilot as a full discover-create-publish pipeline and references approval flows, but it does not clearly warn that autopilot may automatically generate and publish content. In the context of social media account control, autonomous content creation and publication can create reputational, legal, and operational harm if enabled without explicit user understanding and consent.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
97% confidence
Finding

The quick-start example uses autoPublish: true and describes a one-call generate-and-publish flow with no review or confirmation step. This is dangerous because users or downstream agents may copy the example directly, causing immediate publication of AI-generated content to live social accounts without validating accuracy, tone, compliance, or account targeting.

Content

No source excerpt is available for this finding.

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 · references/workflows.md (reported line 8)May include surrounding context.

Every tool call follows this pattern:

bash
curl -X POST https://twinfold.app/api/mcp/tools \
  -H "Authorization: Bearer $TWINFOLD_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"tool": "twinfold.createPost", "arguments": {"topic": "AI trends"}}'

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
96% confidence
Finding

The workflow explicitly recommends createPost with autoPublish: true for a vague user request and presents it as a one-shot autonomous pipeline. In a skill that can post to live social accounts, skipping an explicit confirmation or preview step creates a real risk of unintended public posting, reputational harm, and accidental publication to multiple connected accounts.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Low
Category
Not specified by scanner
Confidence
84% confidence
Finding

The example derives language from connected account settings and uses it to generate and publish content without explicit user opt-in. While not a direct security exploit, it can cause unintended or misleading public output, especially in multilingual environments where account defaults may not match the user's intent for a specific post.

Content

No source excerpt is available for this finding.

Static analysis

No suspicious patterns detected.