The Agent Skills Directory

[PROMPT_INJECTION]: No malicious injection patterns or safety bypass attempts were detected. The instructions focus on structural best practices for agent configuration.
[DATA_EXFILTRATION]: No sensitive file paths, credential patterns, or suspicious network operations were identified. The skill does not contain code to read or transmit data.
[OBFUSCATION]: No hidden content, encoded strings, zero-width characters, or homoglyph attacks were found in the skill files or metadata.
[REMOTE_CODE_EXECUTION]: The skill does not perform any remote script downloads or package installations. It is purely instructional markdown and configuration data.
[PRIVILEGE_ESCALATION]: No attempts to acquire elevated permissions or bypass system security controls were detected. The skill advocates for the principle of least-privilege.
[PERSISTENCE]: No mechanisms for maintaining access, such as shell profile modifications or scheduled tasks, were identified.
[DYNAMIC_EXECUTION]: No use of dynamic code evaluation (e.g., eval, exec) or runtime compilation was found.
[DYNAMIC_CONTEXT_INJECTION]: The skill does not use shell execution patterns within its markdown or configuration files.
[NO_CODE]: The skill contains only documentation, standard operating procedures, and a JSON manifest. No executable scripts are included.

agentsop-tool-scoping