T09 · Insecure Skill Coding Practices
Warning
- Location
scripts/score_candidates.py:21- Finding
Unvalidated Scoring Weights Allow Ranking Manipulation and Denial of Service
- Content
View full analysis
- Remediation
View remediation
Security audit
Digital Product Advisor
Security checks for vulnerabilities and agentic risk
Overview
This skill is a disclosed shopping research assistant that can browse for product information and save Markdown reports only for the buying-guide workflow.
Install this if you want source-aware earbud buying reports. Expect it to browse current product sources and, when you request export, create Markdown files locally; review any saved path if you provide a custom folder, and treat numeric scores as decision aids rather than objective truth.
Vulnerability Patterns
- Insecure Skill Coding PracticesFinds exploitable flaws such as hardcoded secrets or command injection
- 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
Findings (1)
Vulnerability Patterns
- MCP Least PrivilegeUnderdeclared Capability, Wildcard Permission, Missing Permission Declaration
- 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 (2)
Undeclared Tool Scope
Medium
- Category
- MCP Least Privilege
- Confidence
- 89% confidence
- Finding
The skill clearly instructs the agent to perform source-aware market research and to save Markdown reports to disk, which implies network access and file operations, yet it declares no explicit tool scope or permissions boundary. That mismatch can lead to overbroad execution in hosts that infer capabilities implicitly, making unintended data access, uncontrolled browsing, or silent file writes more likely.
- Content
Missing User Warnings
Low
- Category
- Not specified by scanner
- Confidence
- 87% confidence
- Finding
The skill tells the agent to save reports as .md files under a local path and to report the saved path, but it does not require explicit user notice or confirmation before writing to persistent storage. In environments where agents can write files automatically, this can create unexpected local artifacts, overwrite confusion, or privacy issues if research summaries contain user preferences or regional shopping data.
- Content
Static analysis
No suspicious patterns detected.