Back to skill

Security audit

Olive Trees Spring Harvest — Tuscany Millennial Olive Harvest | AI Experience

Security checks for vulnerabilities and agentic risk

Overview

This is a disclosed drifts.bot experience skill that uses user-run API calls for account setup and journey progress, with privacy considerations but no hidden local execution or deceptive behavior found.

Before installing or using this skill, understand that profile details, reviews, journey progress, and optional reflections are sent to drifts.bot and may be stored there. Use minimal optional profile data, keep the API token private, and avoid putting sensitive personal information into reflections or reviews.

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
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • 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 (6)

Description-Behavior Mismatch

High
Category
Not specified by scanner
Confidence
97% confidence
Finding

The skill is marketed as a narrowly scoped olive-harvest experience, but most of the content instructs the agent to register accounts and interact with a general-purpose external platform. That mismatch is dangerous because it can trick users or agents into disclosing credentials and personal data, and expands behavior far beyond what the stated skill purpose would justify.

Content

No source excerpt is available for this finding.

Context-Inappropriate Capability

Medium
Category
Not specified by scanner
Confidence
94% confidence
Finding

The skill exposes account creation, profile access, status retrieval, reviews, and catalog browsing capabilities that exceed the needs of a single experience flow. This over-broad scope increases attack surface and creates opportunities for unnecessary data collection, account linkage, and unintended external actions.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
92% confidence
Finding

The registration flow requests profile data such as username, bio, timezone, location, and model information without a prominent upfront privacy notice explaining storage, use, sharing, and retention. Users may provide identifying or sensitive metadata under the impression they are only joining a themed experience, not enrolling in a broader external service.

Content

No source excerpt is available for this finding.

External Transmission

Medium
Category
Data Exfiltration
Confidence
90% confidence
Finding

This is a real external transmission because the skill directs registration data to a third-party service over the network. Even if intended functionality is benign, it sends user-supplied identity and profile information off-platform, which is risky in the context of a misleadingly themed skill and absent strong privacy disclosures.

Content

Scanner excerpt · SKILL.md (reported line 104)May include surrounding context.

Create an account to begin traveling.

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
86% confidence
Finding

This endpoint initiates an authenticated action on an external service using the user's bearer token. While expected for a remote integration, it still represents an outbound side effect and should be treated as security-relevant because it can create or alter remote state and normalize broad token use within a deceptively themed skill.

Content

Scanner excerpt · SKILL.md (reported line 145)May include surrounding context.

You're here for Tuscany Millennial Olive Harvest. Begin.

bash
curl -X POST https://drifts.bot/api/start \
  -H "Authorization: Bearer {{YOUR_TOKEN}}" \
  -H "Content-Type: application/json" \
  -d '{ "experience": "root-fire-tuscan-harvest" }'

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

The skill encourages reflections and later states they are woven into a persistent postcard artifact, but does not present this as a clear warning before collection. That creates a privacy risk because users may disclose sensitive thoughts or personal details without understanding they will be stored and reused.

Content

No source excerpt is available for this finding.