The Agent Skills Directory

[Prompt Injection] (SAFE): No override markers, bypass instructions, or malicious role-play patterns detected.
[Data Exposure & Exfiltration] (SAFE): No access to sensitive file paths, hardcoded credentials, or network operations found.
[Obfuscation] (SAFE): No encoded strings, homoglyphs, or zero-width characters detected in the content.
[Unverifiable Dependencies & RCE] (SAFE): The skill contains no package installation commands or remote script execution patterns.
[Privilege Escalation] (SAFE): No commands related to system permission modification or administrative access identified.
[Persistence Mechanisms] (SAFE): No attempts to modify shell profiles, system services, or scheduled tasks.
[Metadata Poisoning] (SAFE): Metadata fields are descriptive and align with the skill's purpose.
[Indirect Prompt Injection] (SAFE): The skill serves as a template repository and does not ingest or process untrusted external data.
[Time-Delayed / Conditional Attacks] (SAFE): No logic gating behavior based on time, date, or specific conditions.
[Dynamic Execution] (SAFE): No use of eval, exec, or runtime compilation of code.

documentation-templates