递延所得税与暂时性差异核对（免费版）

• Outcome: Pass
• Latest audit: Sep 20, 2026 · 12:04 PM UTC
• Version: 1.0.55

Overview

This is a local Chinese deferred-tax ledger checker with no network or persistence, though users should read the free-versus-full capability limits carefully.

Install this only if you want a Chinese-language local CLI checker for explicitly supplied deferred-tax workpapers. Treat the free edition as an arithmetic and internal-consistency checker, not a full tax or audit opinion, and do not rely on it for the withheld checks such as tax-rate consistency, loss carryforward eligibility, or income-tax expense tie-out.

A.I.G

Vulnerability Patterns

• Skill Instruction Hijacking: Alters the agent's session goals or safety constraints when the skill loads
• Agent Memory Poisoning: Writes attacker-controlled rules into memory that affect later sessions
• Remote Payload Retrieval and Execution: Fetches external code whose behavior can change after review
• Embedded Malicious Code: Ships malicious scripts inside the skill and executes them locally
• Unauthorized Access and Privilege Escalation: Obtains permissions beyond the task's legitimate needs

SkillSpector

Vulnerability Patterns

• Trigger Abuse: Overly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
• MCP Tool Poisoning: Hidden Instructions, Unicode Deception, Parameter Description Injection
• Prompt Injection: Instruction Override, Hidden Instructions, Exfiltration Commands
• Data Exfiltration: External Transmission, Env Variable Harvesting, File System Enumeration
• Privilege Escalation: Excessive Permissions, Sudo/Root Execution, Credential Access

Findings (9)

Intent-Code Divergence

High

• Confidence: 98% confidence
• Finding: The skill documentation is internally inconsistent: earlier sections advertise checks such as direction/sign consistency and duplicate/blank detection, while the later 'free version does not include' section excludes related or overlapping checks like direction-to-account classification, tax-rate consistency, loss carryforward eligibility, and income-tax expense tie-outs. This can mislead users and downstream agents about what controls actually run, causing reliance on missing validations in a tax/compliance workflow.

Ae1

High

• Category: analysis-evasion

Content

node scripts/run.mjs --sample # 先看样例

• Confidence: 100% confidence
• Finding: Referenced artifact was not completely inspected

Ae1

High

• Category: analysis-evasion

Content

node scripts/run.mjs --sample # 先看样例

• Confidence: 100% confidence
• Finding: Referenced artifact was not completely inspected

Vague Triggers

Medium

• Confidence: 94% confidence
• Finding: The trigger phrases are broad finance terms such as deferred tax, temporary differences, and deferred tax assets/liabilities, which may appear in ordinary discussion rather than an explicit request to invoke the skill. In an agentic environment this can cause accidental activation on unrelated financial content, leading to inappropriate tool use, confusion, or unintended processing of sensitive accounting data.

Natural-Language Policy Violations

Medium

• Confidence: 91% confidence
• Finding: The file’s natural-language documentation and user-facing messages are entirely written in Chinese, and the contract/comments indicate the tool communicates with users in that language by default. For a general-purpose skill, this effectively imposes a locale/language choice without offering an opt-in or alternative, which matches the language-policy violation criterion.

Intent-Code Divergence

Medium

• Confidence: 91% confidence
• Finding: The comments state that withheld checks are only descriptive text and not implemented, but the runtime result deliberately exposes checks_withheld and checks_not_run metadata as part of the engine output. This can misrepresent engine capabilities and state to callers, enabling deceptive behavior in product enforcement, compliance evidence, or user decision-making because the artifact appears aware of premium checks despite claiming they are absent.

Description-Behavior Mismatch

Medium

• Confidence: 89% confidence
• Finding: The file-level contract claims the free package should only execute free checks, yet the source explicitly describes and accommodates a subscription-gated paid branch in the same engine. Even though the paid logic is not obviously executed in this snippet, embedding a paid-feature hook in the free artifact creates a security and trust boundary failure: downstream tooling, build mistakes, or crafted inputs could expose non-free behavior or cause policy bypass around feature separation.

Natural-Language Policy Violations

Medium

• Confidence: 92% confidence
• Finding: The user-facing description, help text, prompts, and error messages are entirely in Chinese, and the file does not offer any alternative language or opt-in choice. Under the policy, forcing a specific language without user opt-in is a natural-language policy violation unless the locale restriction is explicitly documented and justified.

Natural-Language Policy Violations

Low

• Confidence: 90% confidence
• Finding: The only natural-language content in this file is a Chinese-language table embedded in the JSON value. For a generic sample template, this hard-codes a specific language/locale without offering user opt-in or documenting that the skill is region-specific, which matches the policy-violation criteria for locale constraints.

Static analysis

No suspicious patterns detected.