The Agent Skills Directory

[SAFE]: The skill's toolset is limited to read-only operations (Read, Grep, Glob), ensuring it cannot modify files or execute arbitrary system commands.
[SAFE]: No network-related operations (e.g., curl, wget) or external data exfiltration patterns were identified.
[SAFE]: The instructions do not contain any prompt injection, role-play bypasses, or attempts to override the agent's core safety protocols.
[SAFE]: No obfuscation, hidden URLs, or encoded payloads were found in the skill's source or metadata.
[SAFE]: While the skill processes potentially untrusted data (code diffs and file contents), its restricted capabilities prevent any indirect prompt injection from causing meaningful harm, as the agent lacks the tools to perform network or write operations.

predict