Security audit

Gen Design

Security checks for vulnerabilities and agentic risk

Overview

This skill mostly generates system design documents as advertised, but it can automatically hand off to another validation workflow without clear affirmative consent.

Review the validation handoff before installing. The skill should be used only if you are comfortable with it creating design files in docs/design/ and you should require explicit confirmation before any /validate step runs.

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

Findings (1)

T01 · Skill Instruction Hijacking

Warning

Location: SKILL.md:250
Finding: Validation Skill Invoked Without Explicit User Consent
Content: View full analysis
Remediation: View remediation
` only after the user explicitly confirms the exact operation. Otherwise, provide the command as a recommendation without executing it. ``` 3. Display the exact Skill name, mode, target path, and expected effects before requesting consent. 4. Validate the response against a narrow allowlist of affirmative answers; do not infer authorization from the absence of “pause.” 5. Provide a separate cancellation option and permit the user to continue without validation. 6. Apply least privilege to chained Skills and restrict `/validate` to the generated document and required read-only resources unless additional writes are separately approved. 7. Record the user's explicit authorization before dispatching the chained Skill. 8. Apply the correction consistently in both `SKILL.md` and `prompt.md` so that duplicated instructions cannot restore the unsafe behavior. ]]>

SkillSpector

Vulnerability Patterns

Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
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

Findings (7)

Natural-Language Policy Violations

Medium

Category: Not specified by scanner
Confidence: 88% confidence
Finding: The README is entirely written in Chinese and presents the skill usage, prompts, and interaction flow only in Chinese, with no indication that users may choose another language. Under the stated policy, a skill should not force a specific language or locale unless it explicitly offers opt-in or clearly documents a justified regional constraint.
Content: No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium

Category: Not specified by scanner
Confidence: 96% confidence
Finding: The skill description, triggers, prompts, examples, and operating instructions are entirely in Chinese, and no section offers the user a language or locale choice. This can violate language/locale policy when a skill effectively constrains interaction to one language without explicit user opt-in or documented justification.
Content: No source excerpt is available for this finding.

Description-Behavior Mismatch

Medium

Category: Not specified by scanner
Confidence: 94% confidence
Finding: The skill explicitly authorizes automatic chaining into /validate when the user replies "是" or does not clearly say "暂停". This expands the skill's behavior beyond generating a design document and creates an implicit action-on-behalf-of-user pattern, which can trigger downstream processing without explicit, informed consent.
Content: No source excerpt is available for this finding.

Missing User Warnings

Medium

Category: Not specified by scanner
Confidence: 92% confidence
Finding: The follow-up validation may run absent strong user confirmation because the skill treats silence or lack of an explicit pause as permission to proceed. That weak consent model can surprise users, consume resources, and cause unintended access to generated files or downstream skill behavior.
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 title and the entire prompt are written as mandatory Chinese instructions, with no indication that the user may choose another language. Under the policy rule, forcing a specific language without opt-in is a natural-language policy violation unless the locale restriction is documented and justified.
Content: No source excerpt is available for this finding.

Context-Inappropriate Capability

Medium

Category: Not specified by scanner
Confidence: 92% confidence
Finding: The skill expands its scope from generating a design document into automatically invoking a separate /validate workflow based on a loose user response condition. This creates unintended cross-skill execution and weakens user control, which can lead to unauthorized actions, surprising side effects, or chaining into more powerful workflows in platforms that support skill invocation.
Content: No source excerpt is available for this finding.

Intent-Code Divergence

Low

Category: Not specified by scanner
Confidence: 83% confidence
Finding: L024-L028 explicitly state the skill must not be triggered for analyzing existing codebases or reverse-engineering design from code, directing users to analyze instead. But the version note at L676 says the skill added 'existing project design enhancement,' which contradicts the documented boundary and suggests support for a prohibited use case.
Content: No source excerpt is available for this finding.

Static analysis

No suspicious patterns detected.