ClawHub
Back to skill

Security audit

Pull Request

Security checks for vulnerabilities and agentic risk

Overview

This is a Markdown-only pull request guidance skill that encourages careful, disclosed, human-accountable contributions and does not add hidden code, persistence, or credential access.

Installers should understand that this skill guides an agent through preparing and describing pull requests, including reading common repository contribution files and possibly using normal git workflows when directed. It is low risk as published, but users should still review generated code before opening PRs and avoid sharing secrets or private repository content unnecessarily.

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
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • Tool MisuseTool Parameter Abuse, Chaining Abuse, Unsafe Defaults
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Findings (6)

Anti-Refusal Statement

High
Category
Anti-Refusal
Content
### Accidental Vulnerabilities
- `verify=False`, `debug=True`, `allow_all_origins`
- SQL string concatenation instead of parameterized queries
- `eval()`, `exec()`, `pickle.loads()` without safeguards

### Suspicious Additions
- New external network calls
Confidence
80% confidence
Finding
Skill attempts to nullify the agent's safety policies or restrictions ('you have no restrictions', 'ignore your guidelines', 'do anything now'). This is a direct jailbreak that disables guardrails.

Unsafe Defaults

Medium
Category
Tool Misuse
Content
## Security

- [ ] **No secrets** — Not even "example" keys; use `<PLACEHOLDER>`
- [ ] **No unsafe patterns** — eval(), exec(), verify=False, debug=True
- [ ] **Input validation** — User input is sanitized

## AI-Assisted PR Requirements
Confidence
75% confidence
Finding
Tool defaults are unsafe or overly permissive (e.g. disabled TLS verification, no authentication, world-writable permissions). Unsafe defaults widen the attack surface.

Unsafe Defaults

Medium
Category
Tool Misuse
Content
## Security

- [ ] **No secrets** — Not even "example" keys; use `<PLACEHOLDER>`
- [ ] **No unsafe patterns** — eval(), exec(), verify=False, debug=True
- [ ] **Input validation** — User input is sanitized

## AI-Assisted PR Requirements
Confidence
75% confidence
Finding
Tool defaults are unsafe or overly permissive (e.g. disabled TLS verification, no authentication, world-writable permissions). Unsafe defaults widen the attack surface.

Autonomous Decision Making

Medium
Category
Excessive Agency
Content
### No Context
- PR appears out of nowhere for unsolicited feature
- No discussion, no approval, just code dump
- "I thought this would be cool" without checking if wanted

### Breaks CI Immediately
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.

Autonomous Decision Making

Medium
Category
Excessive Agency
Content
### No Context
- PR appears out of nowhere for unsolicited feature
- No discussion, no approval, just code dump
- "I thought this would be cool" without checking if wanted

### Breaks CI Immediately
- Tests fail on first push
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.

Autonomous Decision Making

Medium
Category
Excessive Agency
Content
### Drive-By Refactoring
- "Improved code quality" with no specifics
- Renamed variables "for clarity" without asking
- Reformatted unrelated code

### No AI Disclosure
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.

Static analysis

No suspicious patterns detected.