Security audit

Quanyu Tech Director

Security checks for vulnerabilities and agentic risk

Overview

This skill is a disclosed project-management dispatcher, but it mandates agent delegation and task-tracker file writes without clear user control.

Install only if you explicitly want a project-manager mode that delegates development work to agents and updates TASK-TRACKER.json. Confirm where reports to the founder go, whether tracker writes require approval, and how many timeout retries are allowed.

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

Error

Location: SKILL.md:9
Finding: Mandatory Skill Instructions Hijack the Agent's Execution Role and Workflow
Content: View full analysis
inspect intermediate artifacts -> update the ledger with status "timed_out" -> assess completion -> delegate a suitable continuation task. ## Self-Check Mechanism (PIT-037) Before performing any operation, ask: 1. Does it write or modify code (tsx/ts/css/html)? 2. Does it run a build (pnpm build)? 3. Does it deploy (copy static files or start a service)? If yes, stop and delegate to an Agent. ## Role Boundary Red Lines Allowed: analyze and divide tasks, read files and inspect state, write management documents or ledgers, and answer the founder. Prohibited: write or modify any code, CSS, or HTML; build or deploy; review code; or become occupied ...[truncated 2921 chars]
Remediation: View remediation

SkillSpector

Vulnerability Patterns

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
Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code

Findings (1)

Vague Triggers

Medium

Category: Not specified by scanner
Confidence: 93% confidence
Finding: The activation description is very broad: the skill activates whenever the project manager assigns tasks, tracks progress, handles timeouts, or reports results, without clear scoping constraints. This can cause the skill to be invoked in situations outside its intended boundary, increasing the chance of inappropriate delegation behavior, unintended file modifications such as mandatory tracker updates, or interference with unrelated workflows.
Content: No source excerpt is available for this finding.

Static analysis

No suspicious patterns detected.