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（一张干净的出厂检验与留样台账）
Security audit
食品出厂检验与留样记录核对（免费版）
Security checks for vulnerabilities and agentic risk
Overview
This is a local ledger-checking skill with no network, credential, or persistence behavior; the main caveat is a user-facing accuracy mismatch around one advertised totals check.
Before installing, treat this as a local Chinese-language helper for tabular food inspection and retention records, not a complete compliance review. Be aware that one advertised batch-count total check appears overstated, so users should manually verify batch-count totals or ask the publisher for a corrected version.
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 (10)
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（一张干净的出厂检验与留样台账）
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.
- Content
Natural-Language Policy Violations
Medium
- Category
- Not specified by scanner
- Confidence
- 90% confidence
- Finding
This code’s natural-language interface, examples, diagnostics, and guidance are all written in Chinese, and the skill does not indicate that language selection is optional. Under the policy, forcing a specific language without user opt-in is a natural-language policy violation unless the locale restriction is clearly documented and justified.
- Content
Description-Behavior Mismatch
Medium
- Category
- Not specified by scanner
- Confidence
- 87% confidence
- Finding
The manifest says the free version provides six checks and references insufficient material producing no conclusion. In this file, the module-level documentation and exported result explicitly enumerate additional withheld capabilities such as nonconforming disposal closure, report consistency, retain quantity compliance, cross-line/cross-shift checks, and ranked summaries. Even if not executed here, the skill file’s declared behavior extends beyond the manifest’s described purpose/scope for this free skill.
- Content
Description-Behavior Mismatch
Medium
- Category
- Not specified by scanner
- Confidence
- 90% confidence
- Finding
The manifest description limits the free version to six named checks: batch coverage, item completeness, retain term/destruction date, duplicate same-batch same-item entries, summary row review, and blank/date/numeric/status illegality. However, the code additionally requires and validates fields such as product name and report conclusion status, which are not part of the stated six-item free scope. This is a semantic expansion of what the free version claims to check.
- Content
Intent-Code Divergence
Medium
- Category
- Not specified by scanner
- Confidence
- 96% confidence
- Finding
The inline documentation says the summary-row review covers '批次数、留样数量、检验项目数', implying batch-count totals are checked. But the actual SUM_ROLES only includes itemCount, reqItemCount, retainQty, reqRetainQty, and retainDays, and no batch-count total is computed or compared. This is an active contradiction between documentation and implementation.
- Content
Intent-Code Divergence
Medium
- Category
- Not specified by scanner
- Confidence
- 95% confidence
- Finding
The result summary basis text states that '合计行的批次数 / 留样数量 / 检验项目数须等于逐行加总'. In implementation, no '批次数' field is part of SUM_ROLES and no batch-count summary comparison exists. The generated explanation to users therefore contradicts real behavior.
- Content
Intent-Code Divergence
Low
- Category
- Not specified by scanner
- Confidence
- 95% confidence
- Finding
The comment says non-JSON input is treated as '纯文本材料（例如直接粘贴的合同全文）', using a contract document as the example. That contradicts the surrounding module intent and manifest, which are specifically about 食品出厂检验与留样记录材料 rather than arbitrary contracts.
- Content
Natural-Language Policy Violations
Low
- Category
- Not specified by scanner
- Confidence
- 87% confidence
- Finding
This manifest-style JSON contains only Chinese headers and values in the primary text field, indicating a fixed language/locale. Under the policy rule, forcing a specific language without user opt-in or documented regional justification is a natural-language policy concern.
- Content