Back to skill

Security audit

Automation Workflows

Security checks for vulnerabilities and agentic risk

Overview

This is a non-executable automation playbook whose operational examples are disclosed and aligned with its stated purpose, though users should add privacy and approval safeguards before automating customer or payment workflows.

Before installing or using this skill, treat it as business-process guidance rather than a ready-made secure implementation. For workflows touching customers, payments, public posts, or CRM/accounting records, use least-privilege app permissions, test with non-production data, add approval steps before mass messaging or financial updates, and keep logs plus rollback procedures.

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
  • Rogue AgentSelf-Modification, Session Persistence
  • 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 (3)

Vague Triggers

Medium
Confidence
95% confidence
Finding
The trigger phrases are very broad and include common business language such as 'save time' and 'automation', which can cause the skill to activate in situations where the user did not specifically request workflow automation guidance. Over-broad activation increases the chance of inappropriate invocation, irrelevant advice, or unintended handling of business context that belongs to a different skill.

Missing User Warnings

Medium
Confidence
91% confidence
Finding
The skill recommends automations that move customer, payment, and usage data across multiple systems and even send customer-facing communications, but it does not include safeguards around consent, least-privilege access, data minimization, validation, or rollback for accidental actions. In this context, omissions can lead to privacy violations, corrupted records, mis-segmentation, erroneous emails, or financial workflow mistakes at operational scale.

Session Persistence

Medium
Category
Rogue Agent
Content
2. Flag customers with <50% of average usage
  3. Add flagged customers to "At Risk" segment in CRM
  4. Send re-engagement email campaign to at-risk customers
  5. Create task for me to personally reach out to top 10 at-risk customers
```

### Invoice and payment tracking
Confidence
80% confidence
Finding
Skill establishes unauthorized persistence across sessions via cron jobs, startup scripts, or state files. Session persistence allows an attacker to maintain access beyond the current interaction.

Static analysis

No suspicious patterns detected.