ClawHub
Back to skill

Security audit

Company Intel

Security checks for vulnerabilities and agentic risk

Overview

This skill is a small company-research helper that sends user-provided company context to a local service, with no hidden persistence, privilege changes, or destructive behavior found.

Before installing, make sure the service listening on 127.0.0.1:8010 is one you trust. Avoid sending confidential job descriptions, internal company notes, or private hiring context unless you know how that local service stores data and whether include_search performs external searches.

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
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (3)

Natural-Language Policy Violations

Medium
Confidence
93% confidence
Finding
The activation phrases are entirely in Chinese, which implies a language-specific usage expectation, but the file does not state that the skill is Chinese-only, region-specific, or that users may choose another language. The policy requires flagging language or locale constraints when they are imposed without user opt-in or clear justification.

Missing User Warnings

Medium
Confidence
92% confidence
Finding
The skill transmits user-provided company and role context via an HTTP POST request without any user-facing disclosure or consent step. Even though the destination is localhost, this still sends potentially sensitive job-search or proprietary context to another service, and users are not informed what data leaves the agent context or whether `include_search` may trigger further outbound lookups.

External Transmission

Medium
Category
Data Exfiltration
Content
## Command templates (exec tool + curl)

- Generate company intel:
  - `curl -sS -X POST "http://127.0.0.1:8010/api/company/intel" -H "Content-Type: application/json" -d '{"company":"MiniAgent","role_title":"AI Agent Intern","focus_keywords":["技术栈","面试流程"],"max_results":6,"include_search":true}'`

## Constraints
Confidence
90% confidence
Finding
The command template includes a concrete `curl` POST invocation that sends structured data to a local HTTP service, constituting external transmission from the agent skill to another process. The risk is amplified because the skill encourages operational use of the request without warning about transmitted contents, retention, or whether the backend may enrich data using external search.

Static analysis

No suspicious patterns detected.