The Agent Skills Directory

[PROMPT_INJECTION]: No override instructions, safety bypasses, or system prompt extraction patterns were found in the skill text or metadata.
[DATA_EXFILTRATION]: No network operations, hardcoded credentials, or access to sensitive file paths were detected. The skill focuses on local code manipulation.
[REMOTE_CODE_EXECUTION]: The skill does not perform remote script downloads or execution. It mentions local project-specific validation scripts (distiller.py) in the specification for maintenance and testing.
[COMMAND_EXECUTION]: References to standard version control commands like git blame and git diff are used legitimately within the context of code simplification and understanding logic history.
[OBFUSCATION]: No encoded content, hidden characters, or homoglyph substitutions were identified in the files.
[PRIVILEGE_ESCALATION]: No commands for acquiring elevated permissions (e.g., sudo) or modifying system-level configurations were found.
[INDIRECT_PROMPT_INJECTION]: While the skill processes user-supplied code, its focus is on structural simplification and behavior-preserving refactoring. It does not exhibit dangerous autonomous capabilities that could be easily subverted by adversarial input.

simplifying-code