Back to skill

Security audit

Verifier Skill Generator

Security checks for vulnerabilities and agentic risk

Overview

This skill is mostly coherent, but it asks agents to generate reusable verifier skills with authentication and self-update behavior that is not clearly scoped or user-controlled.

Review generated verifier skills before running or committing them, especially any authentication handling or self-update logic. Prefer requiring explicit approval for updates, pinned templates or versions, and normal code review for generated verifier changes.

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
  • Rogue AgentSelf-Modification, Session Persistence
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Findings (3)

Self-Modification

High
Category
Rogue Agent
Confidence
90% confidence
Finding

The instruction to include 'self-update behavior' in each generated verifier introduces an open-ended self-modification capability into artifacts that will later be discovered and executed by verification tooling. In this context, that makes the generated verifiers harder to audit, increases persistence and supply-chain risk, and could allow future unauthorized changes to execution logic or credentials handling without clear user review.

Content

Scanner excerpt · SKILL.md (reported line 21)May include surrounding context.

md
## Guardrails
- Focus on functional verification, not unit tests or type checks.
- Use names discoverable by downstream verification tooling.
- Include setup, authentication, cleanup, and self-update behavior in each verifier.

## Example Requests
- Generate verifier skills for this frontend and API.

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
89% confidence
Finding

The default prompt is very broad: it tells the agent to inspect an arbitrary project and generate verifier skills for any relevant web, CLI, or API surface. This can cause the skill to activate on generic project-analysis requests and produce powerful derived artifacts from untrusted input, increasing the chance of unintended invocation, overbroad scope, or prompt-driven generation based on adversarial project content.

Content

No source excerpt is available for this finding.

Missing User Warnings

Low
Category
Not specified by scanner
Confidence
84% confidence
Finding

This markdown file states that the skill will 'Generate verifier skill folders in the repo' but does not provide any user-facing warning that it will modify the repository contents. Because creating folders/files can affect user data and workspace state, the description should disclose that behavior clearly before use.

Content

No source excerpt is available for this finding.