Security audit

Results Claim & Hedging Checker

Security checks for vulnerabilities and agentic risk

Overview

This is a Markdown-only academic writing review skill with some scope and language clarity issues, but no evidence of hidden execution, persistence, credential access, or data exfiltration.

Installers should treat this as a narrow academic editing aid, not a statistics validator. Before relying on it, note that statistical-format completeness is delegated elsewhere and that outputs may include Chinese rubric language unless the calling workflow constrains presentation language.

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

SkillSpector

Vulnerability Patterns

Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
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 (9)

Description-Behavior Mismatch

Medium

Category: Not specified by scanner
Confidence: 97% confidence
Finding: The README creates a material inconsistency about scope: earlier it says the skill checks statistical reporting completeness, but the scope boundary later declares statistical reporting format and completeness out of scope and delegated elsewhere. In an agent pipeline, this can cause the skill to silently skip expected checks or produce incomplete reviews, leading users or downstream aggregators to rely on inaccurate coverage assumptions.
Content: No source excerpt is available for this finding.

Description-Behavior Mismatch

Medium

Category: Not specified by scanner
Confidence: 98% confidence
Finding: The skill metadata promises checks for 'statistical reporting completeness,' but the body later excludes statistical reporting format/completeness from scope. This creates a deceptive capability mismatch: users may rely on the skill to catch missing effect sizes, confidence intervals, or reporting omissions that it explicitly does not evaluate, leading to unsafe under-review of academic claims.
Content: No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium

Category: Not specified by scanner
Confidence: 91% confidence
Finding: The skill hard-requires mixed Chinese scoring/output language without asking for the user's language preference, even though inputs may be English. This can cause instruction-following failure, user confusion, and incorrect downstream aggregation if the caller expects output in the user's language or in a stable schema not tied to Chinese prose conventions.
Content: No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium

Category: Not specified by scanner
Confidence: 91% confidence
Finding: The output contract mandates Chinese-language behavioral instructions and rigid heading usage without any opt-in or locale negotiation. In a composable agent system, this is a reliability and integration risk because it can override caller expectations, break parsers, and cause the skill to return unusable output for non-Chinese workflows.
Content: No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium

Category: Not specified by scanner
Confidence: 90% confidence
Finding: The output format section repeatedly enforces Chinese-only instructions and strict heading behavior, increasing the chance that the skill ignores user language, caller policy, or aggregator requirements. This is dangerous in multi-skill pipelines because response-shaping constraints can cause denial of service at the orchestration layer or silently degraded outputs if downstream components cannot parse or present the result correctly.
Content: No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium

Category: Not specified by scanner
Confidence: 95% confidence
Finding: The file’s instructional content is written entirely in Chinese and states usage guidance directly in that language, which effectively imposes a language requirement on users. The checklist later asks whether the source text is English or Chinese, but it does not offer the user a choice of checklist language or explain that this skill is intentionally limited to a Chinese-speaking audience.
Content: No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium

Category: Not specified by scanner
Confidence: 92% confidence
Finding: This markdown file presents the skill’s instructional content in Chinese while the example source text remains in English, and it does not state that the skill is Chinese-only or provide any user language/locale opt-in. That can amount to a language-policy violation because users may be implicitly forced into a specific instructional language.
Content: No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium

Category: Not specified by scanner
Confidence: 96% confidence
Finding: This markdown file contains the substantive skill instructions entirely in Chinese, which effectively forces a specific language for users or downstream agents. Under the policy, language constraints should either be optional/opt-in or clearly justified as region-specific.
Content: No source excerpt is available for this finding.

Natural-Language Policy Violations

Low

Category: Not specified by scanner
Confidence: 78% confidence
Finding: The rubric language is written exclusively in Chinese despite the skill accepting English inputs and being described in English elsewhere. This creates an implicit language constraint that may lead the skill to favor Chinese-language outputs or reasoning without user consent.
Content: No source excerpt is available for this finding.