The Agent Skills Directory

[SAFE]: The skill implements a reasoning workflow for complex problem-solving and architectural analysis. It does not perform network operations, access sensitive credentials, or execute arbitrary shell commands.
[SAFE]: No obfuscation, persistence, or privilege escalation mechanisms were detected. The skill relies on internal reasoning tools and structured prompt templates.
[SAFE]: While the skill ingests user input via the $ARGUMENTS variable, its capability inventory is limited to reasoning through a sequential thinking tool, presenting no significant risk of command or code injection.

oracle-thinkdeep