ClawHub
Back to skill

Security audit

Review Code

Security checks for vulnerabilities and agentic risk

Overview

This is a disclosed, instruction-only code review skill with local memory that requires user confirmation before writing files.

Before installing, decide whether you want it to keep review preferences and finding logs in ~/review-code/. Do not approve saving secrets, credentials, or private code snippets, and keep review scope explicit when asking for PR or merge-readiness checks.

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)

Autonomous Decision Making

Medium
Category
Excessive Agency
Content
- Stored in `~/review-code/`.

**This skill does NOT:**
- auto-approve code or merge pull requests.
- make undeclared network calls.
- store credentials, tokens, or sensitive payloads.
- modify its own core instructions or auxiliary files.
Confidence
85% 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.

Vague Triggers

Medium
Confidence
87% confidence
Finding
The activation guidance is broad enough to trigger on common requests like PR checks, merge readiness, or bug-risk scans, which can cause the skill to engage without explicit user intent. In a code review skill, this creates a scope/consent problem: the agent may steer early conversation toward integration setup instead of directly answering the user, increasing the chance of unsolicited behavior and instruction takeover in normal workflows.

Static analysis

No suspicious patterns detected.