Orchestrated OpenSpec apply — the main agent groups pending tasks by scope and dispatches a worker per group (parallel then serial by dependency), only managing OpenSpec state and reviews; it never writes implementation code itself. Ships /opsx:apply-subagent (Claude subagent worker) with its SOT skill. Requires the agency-agents catalog at ~/.agency-agents (install per the README); a role it cannot resolve there halts the run rather than standing in a fabricated expert.
1 skill · 1 command
Multi-perspective council for one open decision, including whether to keep, replace, or combine an existing draft. It hears conflicting positions before synthesis and ends with a feasible conclusion. The result does not certify the debate. Use review-loop to find and fix defects.
1 skill
Portable visual explainer that turns a complex concept or repository flow into one grounded, self-contained HTML file with large diagrams and few words. Adapted from the Anthropic community ELI5 skill.
1 skill
Black-box user journey review against product goals. Drives the real application, verifies business postconditions, separates failures from environment blocks, and ends with an evidence-based JOURNEY-PASS, JOURNEY-FAIL, or JOURNEY-BLOCKED result.
1 skill
Portable adversarial review loop: Code Reviewer and Reality Checker are always required; the initial round may add zero or multiple relevant domain experts, with no seat cap. It merges findings, applies the smallest accepted fixes, and re-reviews. A repeated blocker gets one fresh escalation before NOT-CONVERGED. Missing catalog or isolation never blocks the loop. Architecture selection belongs to council.
1 skill