The Agent Skills Directory

[SAFE]: The file contains only documentation and process definitions for a software development lifecycle. It does not include any executable scripts, hidden commands, or logic that performs network operations or file system modifications.
[SAFE]: The pipeline design incorporates human-in-the-loop security controls, specifically 'Gate 1' (Plan approval) and 'Gate 2' (Commit approval), which prevent the agent from writing or committing code without explicit user verification.
[SAFE]: The skill explicitly defines a 'Security-review trigger' that mandates involving a specialized security-reviewer agent when the work impacts sensitive areas such as authentication, authorization, cryptography, or secrets management.

orch-pipeline