Back to skill

Security audit

quality-assurance

Security checks for vulnerabilities and agentic risk

Overview

This is a markdown-only quality assurance skill for reviewing other skills, with broad gatekeeper wording but no hidden code, credentials, persistence, or destructive behavior.

Install this if you want an agent-facing checklist for skill QA. Be aware that its description is broad and assertive, so users should invoke it for explicit QA or certification tasks rather than treating its approval as an external publishing authority.

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
  • 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
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Findings (2)

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

The skill description uses very broad, imperative activation language such as validating, testing, or certifying any skill and positioning itself as a gatekeeper. In agent environments that route by semantic similarity, this can cause the skill to be selected in situations where narrower or more appropriate skills should handle the request, leading to unintended execution and overreach.

Content

No source excerpt is available for this finding.

Autonomous Decision Making

Medium
Category
Excessive Agency
Confidence
75% confidence
Finding

Skill enables autonomous high-impact decisions without human-in-the-loop verification. Critical operations (destructive commands, financial transactions, data deletion) should require explicit user confirmation.

Content

Scanner excerpt · references/quality-framework.md (reported line 34)May include surrounding context.

md
### Common Issues
- Step-by-step without thinking
- No verification steps
- No error checking
- Single-path reasoning