Security audit
Security checks for vulnerabilities and agentic risk
This is a prompt-only skill for diagnosing failed AI answers and rewriting retry prompts, with no code execution, network access, credential use, or persistence.
Review the retry prompt before using it and avoid pasting secrets, private records, confidential business data, or sensitive personal information into another AI tool. For medical, legal, financial, safety-critical, or other high-stakes outputs, treat this skill as preparation for independent verification rather than as authoritative advice.
The trigger phrases are broad, conversational, and likely to appear in normal user interactions outside the narrow intended scope of this skill. That increases the chance of unintended activation, which can cause the wrong prompt-flow to run, confuse users, and potentially process sensitive or unrelated AI outputs under an incorrect workflow.
The manifest sets "language": "en", which imposes a specific language/locale in the skill metadata. There is no indication that users may opt into other languages or that the English-only scope is required for a justified region-specific purpose.
No suspicious patterns detected.
