The Agent Skills Directory

[SAFE]: The skill is entirely instructional and does not contain hardcoded credentials, malicious scripts, or commands for unauthorized data exfiltration.
[SAFE]: The skill uses search patterns (e.g., code_search, search) to find information within an organization's repositories. These tools are used according to their intended purpose for repository discovery.
[SAFE]: While the skill involves processing external data (code from repositories) which is a surface for indirect prompt injection, the instructions explicitly guide the agent to perform technical vetting and summarization for a human user, which is a legitimate and safe use case.

find-examples