The Agent Skills Directory

[SAFE]: The skill is primarily instructional and contains no executable code or malicious prompt injection patterns. It provides technical references for the agent to use during code analysis and generation.
[DATA_EXFILTRATION]: No data exfiltration or unauthorized network operations were detected. The skill's inspection of 'tsconfig.json' and local source code is required for providing technical feedback and is consistent with its stated purpose.
[PROMPT_INJECTION]: No direct or indirect malicious instructions were identified. The skill ingests untrusted data such as project configurations and source code (Ingestion points: tsconfig.json and user-provided source files). While it lacks explicit boundary markers or sanitization for this data, it restricts its own capabilities to writing and modifying code (Capability inventory) based on a structured SOP and fixed reference documents, which effectively manages the risk of instruction obedience from user-controlled files.

neo-typescript