Ae1
High
- Category
- analysis-evasion
- Confidence
- 100% confidence
- Finding
Referenced artifact was not completely inspected
- Content
md node scripts/run.mjs --sample # 先看样例
Security audit
材料领用与定额损耗核对（免费版）
Security checks for vulnerabilities and agentic risk
Overview
This skill is a local, Chinese-language table-checking utility with no network use or persistence, though users should note its free version does not perform every advertised accounting check consistently.
Install only if you are comfortable using a Chinese-language local CLI for material usage/loss table checks. Treat the free version as limited: rely on the printed checks_given/checks_not_run lists, and do not assume it verifies amount equals issued quantity times unit price despite the broader disclaimer wording.
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
Vulnerability Patterns
- MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
- Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
- Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
- Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
- Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (7)
Ae1
High
- Category
- analysis-evasion
- Confidence
- 100% confidence
- Finding
Referenced artifact was not completely inspected
- Content
md node scripts/run.mjs --sample # 先看样例
Intent-Code Divergence
Medium
- Category
- Not specified by scanner
- Confidence
- 84% confidence
- Finding
Lines L55-L63 assert that this free version does not include several specific checks and that the only differences from the full version are those listed. But L36 says the actual executed checks are determined by the engine's CHECKS_* exports and should be inspected via --sample, implying the markdown may not be authoritative. This is intent/documentation divergence because the file simultaneously presents a fixed exclusion list and defers truth to code-driven exports.
- Content
Natural-Language Policy Violations
Medium
- Category
- Not specified by scanner
- Confidence
- 94% confidence
- Finding
The file’s comments, user guidance, error messages, and result text are written entirely in Chinese, including strings returned to users in
insufficient, findings messages, and the finalnote/disclaimer. There is no opt-in, language selection mechanism, or documented justification that this skill is intentionally limited to a Chinese-speaking or region-specific context.- Content
Intent-Code Divergence
Medium
- Category
- Not specified by scanner
- Confidence
- 97% confidence
- Finding
The disclaimer states the tool checks the arithmetic relation “issued quantity × unit price = amount,” but the free-version logic never performs that validation. In a financial/material-control skill, this can mislead users into trusting that amount fields were verified when manipulated or erroneous totals would pass unchecked, creating a silent integrity gap.
- Content
Natural-Language Policy Violations
Low
- Category
- Not specified by scanner
- Confidence
- 97% confidence
- Finding
The user-facing natural language throughout the script, including the title, usage text, help output, errors, and operational notes, is entirely in Chinese. There is no indication that users may choose another language or locale, which can violate a language/locale policy when a skill mandates one language by default without opt-in.
- Content
Intent-Code Divergence
Low
- Category
- Not specified by scanner
- Confidence
- 95% confidence
- Finding
The comment on L075 says invalid JSON is handled as plain-text material such as a '合同全文' (full contract text). The surrounding code and manifest are specifically for material usage/loss table checking, and the fallback path on L095 labels the same input as '纯文本单证材料', so the documentation text is inconsistent with the implemented skill intent.
- Content