ClawHub
Back to skill

Security audit

SQL Query Reviewer

Security checks for vulnerabilities and agentic risk

Overview

This is a simple SQL review skill with no executable code, but it should be used carefully around private SQL because it asks the agent to remember aggregate review patterns.

Install only if you want an agent to review SQL text you provide. Avoid pasting secrets, production credentials, or highly sensitive schema details, and consider removing or ignoring the self-improvement tracking instruction if you do not want aggregate patterns remembered between reviews.

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
  • 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 'Use when' section lists short generic triggers like 'optimize this query', 'why is my query slow', and 'check my SQL', plus a very broad condition of 'Any raw SQL query, ORM-generated SQL, or migration file'. In a markdown skill description, these phrases do not define clear activation boundaries or exclusion conditions, which could cause unintended invocation for general troubleshooting requests that merely mention SQL.

Content

No source excerpt is available for this finding.

Context-Inappropriate Capability

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

The skill instructs itself to track findings across reviews and only surface aggregate results after 20 reviews, which implies cross-session state retention unrelated to the core task of reviewing a single SQL query. In an agent environment, this can encourage persistence of user-derived data or hidden memory accumulation without clear consent, creating privacy and scope-control risks even if no explicit storage mechanism is shown.

Content

No source excerpt is available for this finding.