Ae1
High
- Category
- analysis-evasion
- Content
node scripts/run.mjs --sample # 先看样例（一张算得平的干净表，用来确认格式）
- Confidence
- 100% confidence
- Finding
- Referenced artifact was not completely inspected
Security audit
仓储费与库龄结算核对（免费版）
Security checks for vulnerabilities and agentic risk
Overview
This skill is a local warehouse-fee table checker; its artifacts show no network, persistence, credential use, or data mutation, though some documentation is inconsistent.
Install only if you are comfortable using a Chinese-language local CLI checker for warehouse fee tables. Treat the output as an accounting review aid, not payment authorization. The publisher should clean up the contradictory telemetry, --no-stats, and free-versus-full documentation so users can rely on the stated behavior.
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 Least PrivilegeUnderdeclared Capability, Wildcard Permission, Missing Permission Declaration
- 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
Findings (13)
Ae1
High
- Category
- analysis-evasion
- Content
node scripts/run.mjs --sample # 先看样例（一张算得平的干净表，用来确认格式）
- 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（一张算得平的干净表，用来确认格式）
- Confidence
- 100% confidence
- Finding
- Referenced artifact was not completely inspected
Lp1
High
- Category
- MCP Least Privilege
- Confidence
- 75% confidence
- Finding
- The skill uses 'env' capability that is not listed in its permissions. This may indicate deceptive intent or missing permission declarations.
Natural-Language Policy Violations
Medium
- Confidence
- 87% confidence
- Finding
- The main user-facing instructions, usage guidance, outputs, and warnings in the markdown body are written entirely in Chinese, while only a brief English metadata description is provided. This can amount to forcing a specific language for operation without explicit user opt-in or an offered language choice.
Intent-Code Divergence
Medium
- Confidence
- 97% confidence
- Finding
- Lines L002-L003 explicitly say this file is a free-tier subset and that the full paid implementation is not in this package. But the later header at L007 labels this file as "warehouse-storage-fee-check-full.js" and describes the complete edition, which actively contradicts the earlier documentation and the file's actual role.
Natural-Language Policy Violations
Medium
- Confidence
- 95% confidence
- Finding
- The file’s instructional comments and user-facing strings are entirely in Chinese, including guidance, warnings, findings, and returned advice. There is no indication that the skill is region-specific or that users can choose a language, which can violate a language/locale policy requiring user choice or documented justification.
Description-Behavior Mismatch
Medium
- Confidence
- 98% confidence
- Finding
- The file documentation states "免费档执行 8 项" while the manifest description says the free edition executes 6 free checks. The implemented run path also performs more than six checks, so the skill behavior/documentation does not match the manifest's stated scope.
Natural-Language Policy Violations
Medium
- Confidence
- 93% confidence
- Finding
- The file’s natural-language interface, help text, status messages, and usage guidance are entirely in Chinese, and there is no indication that users can opt into another language or locale. This can violate language/locale policy requirements when a skill imposes a single language without user choice or documented regional justification.
Intent-Code Divergence
Medium
- Confidence
- 98% confidence
- Finding
- The script repeatedly claims that it performs anonymous usage telemetry and that users can disable it with --no-stats, but the implementation shown contains neither telemetry logic nor parsing for --no-stats. This is a security-relevant integrity issue because users are being asked to trust privacy and network-behavior statements that do not match the executable interface, making informed consent and downstream review harder.
Intent-Code Divergence
Low
- Confidence
- 95% confidence
- Finding
- The disclaimer says "超收基准取的是本表同款物料的单价均值", implying the tool computes an overcharge baseline from average unit prices for the same material. No such logic exists in the implemented checks; the code only performs formula recomputation, totals, duplicate/blank/negative/integer checks.
Intent-Code Divergence
Low
- Confidence
- 87% confidence
- Finding
- The inline comment says invalid JSON is treated as plain-text material such as 'directly pasted full contract text', but the skill is described as checking warehouse storage fee statements and aging-detail tables. That comment documents a materially broader document type than the manifest and surrounding usage text, indicating intent/documentation drift.
Natural-Language Policy Violations
Low
- Confidence
- 92% confidence
- Finding
- The JSON value consists entirely of Chinese-language field labels and data formatting, which indicates a fixed locale/language output. Because there is no accompanying natural-language indication of user opt-in or documented region-specific justification in this file, it may violate the policy against forcing a specific language without choice.
Static analysis
No suspicious patterns detected.