The Agent Skills Directory

[Data Exposure & Exfiltration] (SAFE): The scripts/project_analyzer.py script reads local package.json files to identify frameworks. It contains no network operations, hardcoded credentials, or exfiltration logic.
[Remote Code Execution] (SAFE): No remote scripts are downloaded or executed. The provided Python script uses standard libraries and performs static file checks.
[Prompt Injection] (SAFE): No malicious override instructions or bypass attempts were found in the markdown documentation or metadata.
[Obfuscation] (SAFE): All files contain clear, readable text and code with no encoding or hidden characters detected.
[Indirect Prompt Injection] (LOW): The project analyzer reads external package.json files. While this is an ingestion point for untrusted data, the script only extracts specific JSON keys for comparison and does not interpolate the content into prompts or execute it, minimizing the risk surface.

core-development