Back to skill

Security audit

Brainstorming Duhy40

Security checks for vulnerabilities and agentic risk

Overview

This is a planning workflow skill that reads project context and creates a committed design spec, with no evidence of hidden code, credential access, or data exfiltration.

Install this only if you want a design-first workflow that may inspect the current repository, ask staged questions, create and commit a spec file, and then move into implementation planning. Review the broad activation wording if you do not want routine small edits routed through this process.

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
  • Rogue AgentSelf-Modification, Session Persistence
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • 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 (3)

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

The skill description is extremely broad ('starting any creative work', 'building components', 'adding functionality', 'modifying behavior'), so it can be auto-invoked for a large fraction of ordinary development requests. Over-broad routing is dangerous because it can hijack user intent, force unnecessary workflow steps, and steer sessions into file inspection, task creation, commits, and downstream skill invocation without a narrowly justified trigger.

Content

No source excerpt is available for this finding.

Session Persistence

Medium
Category
Rogue Agent
Confidence
87% confidence
Finding

The instruction to create and persist tasks for every checklist item pushes the agent to maintain session state and follow a rigid multi-step workflow across the interaction. This becomes risky in combination with instructions to inspect project files, review recent commits, write a design doc to disk, commit it, and invoke another skill, because it can cause unnecessary retention and propagation of context beyond the user's immediate request.

Content

Scanner excerpt · SKILL.md (reported line 22)May include surrounding context.

md
## Checklist

You MUST create a task for each of these items and complete them in order:

1. **Explore project context** �?check files, docs, recent commits
2. **Offer the visual companion just-in-time** �?NOT upfront. The first time a question would genuinely be clearer shown than described, offer it then (its own message); on approval its browser tab opens for you. If no visual question ever arises, never offer it. See the Visual Companion section below.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
98% confidence
Finding

Lines 36-39 contain Chinese instructions and approval criteria in a document otherwise written in English. This imposes a language expectation without user opt-in and could prevent some users or operators from understanding mandatory process requirements.

Content

No source excerpt is available for this finding.