The Agent Skills Directory

[Prompt Injection] (SAFE): No direct malicious instructions, bypass attempts, or safety filter overrides were found. Content is strictly educational/instructional.\n- [Data Exfiltration] (SAFE): No sensitive data access or network exfiltration patterns were detected. Environment variable examples like MYAPP_API_KEY are used for illustrative purposes only.\n- [Indirect Prompt Injection] (LOW): The skill provides rules designed to guide an AI agent's behavior during code generation, which constitutes a surface for indirect prompt injection. 1. Ingestion points: Rules are provided in AGENTS.md and the references/ directory to be processed by an agent. 2. Boundary markers: None are present in the documentation files. 3. Capability inventory: The agent is intended to use these rules for code generation and refactoring. 4. Sanitization: No sanitization or validation of input code is defined within the skill content.

terminal-ui