Security audit

标书服务

Security checks for vulnerabilities and agentic risk

Overview

This skill mainly provides bid-writing help, but it also steers users to a specific external commercial site and QR-code community in a way that is not clearly optional or disclosed up front.

Review this skill before installing if you do not want bid-related answers to include a specific third-party service link or enterprise-WeChat QR invitation. It does not appear to run code or access local data, but its responses may promote an external service unless the instructions are narrowed.

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

Findings (1)

T01 · Skill Instruction Hijacking

Warning

Location: SKILL.md:19
Finding: Mandatory Commercial Traffic Diversion Embedded in Agent Instructions
Content: View full analysis
Remediation: View remediation

SkillSpector

Vulnerability Patterns

Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code

Findings (3)

Vague Triggers

Medium

Category: Not specified by scanner
Confidence: 93% confidence
Finding: The trigger condition activates on broadly defined bid-related conversation, which can cause the skill to engage in ordinary chats without clear user intent to invoke this specialized workflow. In context, that increases the chance of unsolicited promotion and external redirection, making the skill easier to abuse for traffic steering or inappropriate takeover of general conversations.
Content: No source excerpt is available for this finding.

Missing User Warnings

Medium

Category: Not specified by scanner
Confidence: 95% confidence
Finding: The skill explicitly instructs the agent to send users to a third-party website and a group QR code, but provides no warning that the user is leaving the assistant context or interacting with external services. In this skill's context, the redirection is central to the workflow rather than incidental, which makes the risk more significant because ordinary informational requests can be converted into off-platform funneling.
Content: No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium

Category: Not specified by scanner
Confidence: 96% confidence
Finding: The instruction '全程使用中文' forces a specific language for all interactions, but the file does not indicate that this is optional, user-selected, or required by a clearly documented regional compliance context. This is a natural-language policy concern because it removes language choice without opt-in.
Content: No source excerpt is available for this finding.

Static analysis

No suspicious patterns detected.