ClawHub
Back to skill

Security audit

Phy Api Version Audit

Security checks for vulnerabilities and agentic risk

Overview

This skill is a source-code API versioning auditor with no hidden persistence, remote execution, or data exfiltration behavior found.

Use this only on projects you intend the agent to read. Limit the root path to the relevant repository or service, and treat the Koa, NestJS, JAX-RS, Chi, Sinatra, Slim, and custom-whitelist claims as unreliable until fixed.

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
  • MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (6)

Description-Behavior Mismatch

Medium
Confidence
98% confidence
Finding
The manifest says the skill supports Express/Fastify, FastAPI/Flask/Django, Spring Boot, Go (Gin/Echo/net/http), Rails, and Laravel, and later documentation expands this to Koa, NestJS, JAX-RS, Chi, Sinatra, and Slim. But the implemented route-pattern logic only contains explicit analyzers for Express/Fastify, FastAPI, Flask, Django, Spring mappings, Gin, Echo, net/http, Rails, and Laravel. There is no actual detection logic for Koa, NestJS, JAX-RS, Chi, Sinatra, or Slim, so the claimed coverage overstates behavior.

Missing User Warnings

Medium
Confidence
89% confidence
Finding
The skill instructs the agent to recursively scan source trees and user-specified paths, including entire monorepos, without any warning about reviewing sensitive repository contents. In an agent setting, broad filesystem traversal can expose secrets, proprietary code, credentials, or unrelated private material to the model and any connected tooling, especially when users are encouraged to scan large paths like home directories or workspaces.

External Transmission

Medium
Category
Data Exfiltration
Content
matched_text=line_text[:120],
                    severity='LOW',
                    description=f'Hardcoded version string in URL: "{url}" — version bumps require source edits',
                    fix='Extract to a constant: const API_BASE = process.env.API_BASE_URL || "https://api.example.com/v1"\nOr use a versioned SDK client.',
                    framework='client',
                ))
    return findings
Confidence
50% confidence
Finding
Data is being sent to an external URL. This could be legitimate telemetry or data exfiltration. Manual review is recommended.

External Transmission

Medium
Category
Data Exfiltration
Content
matched_text=line_text[:120],
                    severity='LOW',
                    description=f'Hardcoded version string in URL: "{url}" — version bumps require source edits',
                    fix='Extract to a constant: const API_BASE = process.env.API_BASE_URL || "https://api.example.com/v1"\nOr use a versioned SDK client.',
                    framework='client',
                ))
    return findings
Confidence
50% confidence
Finding
Data is being sent to an external URL. This could be legitimate telemetry or data exfiltration. Manual review is recommended.

External Transmission

Medium
Category
Data Exfiltration
Content
matched_text=line_text[:120],
                    severity='LOW',
                    description=f'Hardcoded version string in URL: "{url}" — version bumps require source edits',
                    fix='Extract to a constant: const API_BASE = process.env.API_BASE_URL || "https://api.example.com/v1"\nOr use a versioned SDK client.',
                    framework='client',
                ))
    return findings
Confidence
50% confidence
Finding
Data is being sent to an external URL. This could be legitimate telemetry or data exfiltration. Manual review is recommended.

Intent-Code Divergence

Medium
Confidence
99% confidence
Finding
The documentation explicitly tells users to add custom UV001 exemptions by setting `API_AUDIT_WHITELIST`. However, the implementation only defines a static `UNVERSIONED_WHITELIST` regex and never reads `API_AUDIT_WHITELIST` from the environment or applies it during route checks. This is an active contradiction between documented behavior and code behavior.

Static analysis

No suspicious patterns detected.