← ClaudeAtlas

graphlisted

Graph Engineering orchestrator. Decomposes a task into a work graph of agent/validator/merge/human nodes (DAG) and manages its execution with context isolation, retry gates, and run logs. Trigger on /eff:graph or $eff:graph, and on natural-language asks like: 'graph engineering', 'work graph', 'orchestrate this as a graph', 'split into parallel agents', 'multi-agent orchestration', 'fan out agents with a validation gate', '그래프 엔지니어링', '그래프로 설계해줘/분해해줘', '에��전트 병렬로 나눠서', '역할 나눠서 병렬 진행', '검증 게이트 넣어서 진행' — or any task asking to coordinate multiple agent roles with validation or human-approval gates. Keep this distinct from eff:blitz, which force-parallelizes overlapping implementation work.
im-ian/efficiency · ★ 0 · AI & Automation · score 65
Install: claude install-skill im-ian/efficiency
Read `PROMPT.md` in this skill directory completely and adopt it as the operating instructions. Treat the text following the skill invocation as the task. Apply the layer decision first (prompt → loop → graph); build a graph only when the task genuinely needs one. If the task contains `PLAN ONLY`, stop after the Mermaid diagram and do not execute. After emitting graph JSON, validate it by saving the JSON to a temporary file and running `python3 <this-skill-directory>/scripts/test.py <file>`. Fix the graph until the validator prints `OK` before executing it.