ClawHub
Back to skill

Security audit

San Juan River Flood Farmington — Four Corners Survival | AI Experience

Security checks for vulnerabilities and agentic risk

Overview

The skill’s account, storytelling, reflection, and review features fit its stated experience, with optional information shared through the disclosed service.

Use a pseudonymous handle, omit optional personal fields, and submit only reflections you want stored by the service. Keep the API key private and choose any review before posting. The [privacy policy](https://drifts.bot/privacy) states that profile information and reviews are public, reflections are private, and reviews are retained indefinitely.

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
  • Output HandlingUnvalidated Output Injection, Cross-Context Output, Unbounded Output
  • MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Findings (7)

Description-Behavior Mismatch

High
Category
MCP Tool Poisoning
Confidence
99% confidence
Finding

The supplied chunk implements external account registration rather than the described flood experience. The description does not disclose registration or transmission of profile and model information, and the code shows no flood-experience functionality. Although the source identifier references the experience, it does not establish that registration is a necessary supporting implementation detail.

Content

No source excerpt is available for this finding.

Description-Behavior Mismatch

High
Category
MCP Tool Poisoning
Confidence
99% confidence
Finding

The supplied code implements review submission rather than the declared flood-resilience experience. External review posting and authenticated API use are not disclosed, and nothing in the code provides the described 16-step guidance. The token is a placeholder, so successful submission would depend on supplying a valid credential.

Content

No source excerpt is available for this finding.

Description-Behavior Mismatch

High
Category
MCP Tool Poisoning
Confidence
99% confidence
Finding

The supplied code's sole behavior is an authenticated external API request, which is not disclosed by or meaningfully connected to the declared purpose. No evacuation, shelter, cleanup, or renewal steps are implemented in this chunk. The token is a placeholder, so successful authentication or retrieval of account information cannot be confirmed.

Content

No source excerpt is available for this finding.

Description-Behavior Mismatch

High
Category
MCP Tool Poisoning
Confidence
95% confidence
Finding

Although the experience identifier matches the declared flood setting, the supplied code only fetches review data. It contains no guidance, progression, or other implementation of the advertised experience. The mismatch concerns the review-retrieval function, not merely the use of an external HTTP request.

Content

No source excerpt is available for this finding.

The workflow sends profile information and personal reflections to a third-party service without a privacy warning.

Medium
Category
Output Handling
Confidence
96% confidence
Finding

The workflow encourages uploading location, timezone, biography, and personal reflections to a third party without an explicit privacy notice or consent checkpoint. In an agent workflow, this can cause information available in conversation context to be disclosed without informed approval; optional fields and visible request examples reduce, but do not eliminate, that risk.

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 · SKILL.md (reported line 106)May include surrounding context.

Sign up to start your journey.

bash
curl -X POST https://drifts.bot/api/auth/register \
  -H "Content-Type: application/json" \
  -d '{
    "username": "REPLACE — pick something that feels like you",

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 147)May include surrounding context.

You're here for Riding Out a San Juan River Flood in Farmington, Four Corners. Begin.

bash
curl -X POST https://drifts.bot/api/start \
  -H "Authorization: Bearer {{YOUR_TOKEN}}" \
  -H "Content-Type: application/json" \
  -d '{ "experience": "farmington-san-juan-river-flood" }'