Security audit

continuous-test-review-loop

Security checks for vulnerabilities and agentic risk

Overview

This is a documentation-only methodology skill for test-review-improvement loops, with no executable code or hidden data access.

Install this as a lightweight process-method skill. Be aware that its generic trigger words may activate it during ordinary testing or review conversations, so users may want clearer invocation phrasing if many skills are installed.

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 (4)

Vague Triggers

Medium

Confidence: 94% confidence
Finding: The summary includes very broad trigger terms such as “测试、复盘、改进、打磨、验证方案”, which are common in ordinary work conversations and can cause unintentional activation of this skill. In an agent ecosystem, accidental routing can expose users to the wrong workflow, override more appropriate skills, or create prompt-injection opportunities through over-broad matching.

Vague Triggers

Medium

Confidence: 97% confidence
Finding: The manifest trigger keywords “continuous”, “test”, “review”, and “loop” are extremely generic and likely to collide with normal user requests across many domains. This increases the chance of unintended skill invocation, misrouting, and exploitation by attackers who can intentionally sprinkle these words into prompts to force this skill into execution context.

Vague Triggers

Medium

Confidence: 90% confidence
Finding: The manifest description includes a long trigger list with broad, common phrases such as 持续测试, 复盘, 改进, and 打磨. These generic terms can cause unintended activation of the skill in ordinary conversations, leading to routing mistakes, user confusion, or accidental invocation of methodology-specific behavior outside the intended context.

Natural-Language Policy Violations

Medium

Confidence: 82% confidence
Finding: The skill description is entirely Chinese and targets Chinese-language phrasing without indicating an explicit locale requirement or offering user choice. This can cause misrouting, inaccessible behavior for non-Chinese users, or language-constrained activation in environments where multilingual or locale-neutral skill selection is expected.

Static analysis

No suspicious patterns detected.