Security audit

data-rights-guard

Security checks for vulnerabilities and agentic risk

Overview

This is a simple training-data license checker with documentation inconsistencies, but I found no hidden network access, persistence, or destructive behavior.

Before installing, treat this as a lightweight advisory checker, not a complete legal or security control. Verify the CLI path and arguments in your runner, narrow auto-triggering if broad compliance prompts are a problem, and align the manifest permissions/risk rationale with the actual no-network, no-dependency script behavior.

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
Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
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

Findings (9)

Intent-Code Divergence

High

Category: Not specified by scanner
Confidence: 98% confidence
Finding: The risk rationale says the skill may perform API calls, local execution, and file read/write activity, while the manifest explicitly declares network: false, filesystem: false, and subprocess: false. This contradiction undermines trust boundaries and can lead operators to authorize or deploy the skill under incorrect assumptions, potentially allowing data handling or execution behavior outside the reviewed permission model.
Content: No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium

Category: Not specified by scanner
Confidence: 80% confidence
Finding: The file title and all instructional content are presented in Chinese, and no language-choice or opt-in mechanism is described for users or maintainers. Under the stated policy, forcing a specific language without user opt-in is a natural-language policy concern unless the locale constraint is justified.
Content: No source excerpt is available for this finding.

Vague Triggers

Medium

Category: Not specified by scanner
Confidence: 89% confidence
Finding: The skill’s activation text is very broad and conversational, covering many common questions about data use, training legality, and commercial rights without tight gating criteria. In an agentic system, this can cause over-triggering on loosely related user prompts, leading the compliance skill to intercept workflows unnecessarily or influence decisions outside its intended scope.
Content: No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium

Category: Not specified by scanner
Confidence: 90% confidence
Finding: The policy block hard-codes a country/locale framing (e.g., 'China') and presents a required unified wording block to be copied intact, rather than offering language or locale choice. This can conflict with organizational language/locale policies because the skill appears to mandate a specific locale-specific presentation without user opt-in or a clearly justified regional-only scope.
Content: No source excerpt is available for this finding.

Intent-Code Divergence

Medium

Category: Not specified by scanner
Confidence: 93% confidence
Finding: The manifest and usage text claim active execution through scripts/rights_guard.py / rights_guard.py, but the provided package content here contains only documentation and no corresponding executable artifact. This is dangerous because users or orchestration systems may rely on security/compliance checks that do not actually exist, creating a false sense of protection and causing unverified data-ingestion decisions to be treated as enforced controls.
Content: No source excerpt is available for this finding.

Vague Triggers

Medium

Category: Not specified by scanner
Confidence: 85% confidence
Finding: The markdown says the skill applies to 'any question touching data ownership' and supports automatic routing from natural-language queries containing broad phrases like 'data license' or 'training compliance'. Without exclusion conditions or a narrower context boundary, these triggers could overlap with ordinary compliance discussion and cause unintended invocation.
Content: No source excerpt is available for this finding.

Missing User Warnings

Medium

Category: Not specified by scanner
Confidence: 78% confidence
Finding: The workflow states that data intake is 'blocked automatically' when authorization is missing, but the user-facing description does not clearly foreground that execution may make or enforce gating decisions. Hidden or insufficiently disclosed automatic blocking can disrupt pipelines, cause unexpected denial of service for data ingestion, and make operators unaware that the skill can influence workflow outcomes.
Content: No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium

Category: Not specified by scanner
Confidence: 93% confidence
Finding: The file states in Chinese that this document is the AI-readable skill index and gives no indication that other languages are supported. Because language/locale policy violations apply to all file types, a skill-facing instruction that effectively fixes the language without user opt-in should be flagged.
Content: No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium

Category: Not specified by scanner
Confidence: 95% confidence
Finding: This file contains user-facing natural-language content entirely in Chinese, including the module description and all CLI output messages. Under the policy, forcing a specific language without user opt-in is a natural-language policy violation unless the locale constraint is clearly documented and justified.
Content: No source excerpt is available for this finding.