The Agent Skills Directory

[SAFE]: No malicious patterns, obfuscation, or unauthorized behaviors were found in the skill definitions.\n- [COMMAND_EXECUTION]: The workflow relies on code search and file reading to discover architectural patterns. It does not attempt to execute arbitrary shell commands or system binaries.\n- [DATA_EXFILTRATION]: The skill contains no network-related instructions (e.g., curl, wget) and strictly adheres to writing output within the local repository root.\n- [PROMPT_INJECTION]: While the skill analyzes untrusted project files which provides a surface for indirect prompt injection, its instructions are narrowly focused on generating documentation and diagrams, minimizing the impact of potential injection.

analyse-architecture