Back to skill

Security audit

TestRight: Get Answers Without Endless Testing

Security checks for vulnerabilities and agentic risk

Overview

TestRight provides scoped testing guidance and an optional local helper, with no evidence of hidden access or harmful behavior.

Choose the testing target and set execution, data and spending limits. The optional helper saves copies of experiment inputs locally, so protect sensitive test material. Its consistency checks do not guarantee complete runtime isolation.

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
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • MCP Least PrivilegeUnderdeclared Capability, Wildcard Permission, Missing Permission Declaration
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (2)

Skill Enumeration

Medium
Category
Agent Snooping
Confidence
80% confidence
Finding

Skill enumerates or reads other installed skills. Access to other skills' SKILL.md files or the skills directory reveals prompt instructions, capabilities, and secrets that should be invisible to peer skills.

Content

Scanner excerpt · REFERENCES.md (reported line 11)May include surrounding context.

md
The author observed GPT 6.1 Sol adding excessive instructions to a "before" baseline case, invalidating comparison with the "after" case. The ordinary user request was replaced with a coached task.

OpenAI's [Maintaining software with the GPT-6 family](https://developers.openai.com/blog/maintaining-software-with-the-gpt-6-family), October 14, 2026, recommends replaying the original session prompt from the same repository snapshot and model while changing the instruction under investigation. Anthropic's [skill-creator](https://github.com/anthropics/skills/blob/main/skills/skill-creator/SKILL.md) likewise uses the same task prompt with and without a new skill.

**Instruction:** Preserve the actual user request and relevant starting context. If no original request exists, construct the shortest ordinary request for that use. Apply the evaluated change only in the experimental run. Keep evaluator-added methods outside baseline inputs.

Skill declares no tool scope ('permissions' or 'allowed-tools') but code capabilities were detected: file_read, file_write.

Medium
Category
MCP Least Privilege
Confidence
70% confidence
Finding

Without declared permissions the skill's intent is opaque and cannot be validated.

Content

No source excerpt is available for this finding.