Back to skill

Security audit

Backend Interview Simulator

Security checks for vulnerabilities and agentic risk

Overview

This skill is a backend interview coach that stores limited local practice progress and uses optional CV input, with no evidence of hidden network, destructive, or privileged behavior.

Install this if you want a backend interview practice assistant that keeps local progress files. Be aware it may activate on broad phrases like "system design" unless your request is clearly outside backend topics, and only provide CV files you intend it to read and summarize into its local profile data.

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
  • 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
Findings (5)

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
93% confidence
Finding

The trigger phrase "design interview" is broad enough to activate on requests outside backend scope, causing the skill to take over conversations it was not intended to handle. In an agent environment, overly broad activation can lead to incorrect tool use, irrelevant persistence, or mishandling of user data under the wrong skill context.

Content

No source excerpt is available for this finding.

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

The phrase "Rate my solution" is generic and could match many unrelated tasks, leading the skill to activate on non-backend content. This increases the risk of context confusion, accidental file handling, and storing interview-related state for requests that do not belong to this skill.

Content

No source excerpt is available for this finding.

Session Persistence

Medium
Category
Rogue Agent
Confidence
60% confidence
Finding

Skill establishes unauthorized persistence across sessions via cron jobs, startup scripts, or state files. Session persistence allows an attacker to maintain access beyond the current interaction.

Content

Scanner excerpt · SKILL.md (reported line 44)May include surrounding context.

On first message, ensure data directory exists:

bash
mkdir -p ~/.openclaw/backend-interview-simulator

Initialize (create if missing) using these shapes:

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
88% confidence
Finding

The standalone trigger "system design" lacks a backend qualifier, so the skill may activate for frontend, product, hardware, or general architecture discussions. Misrouting requests to this skill can cause inappropriate questioning, faulty guidance, or unnecessary local persistence under an incorrect user workflow.

Content

No source excerpt is available for this finding.

Intent-Code Divergence

Low
Category
Not specified by scanner
Confidence
78% confidence
Finding

Lines L078-L080 state that storage should be limited to ~/.openclaw/backend-interview-simulator/ only. However, L230 instructs the skill to back up a corrupted file without specifying that the backup must remain in that same directory, which weakly contradicts the earlier documented storage boundary.

Content

No source excerpt is available for this finding.

Static analysis

No suspicious patterns detected.