prompts.chatprompts.chat
Login
All teams

DAG Orchestra

Decompose → pack context → collect results for complex work that needs a dependency graph, not one long chat. Inspired by orchestration plugins that dispatch token-budgeted subagents on a DAG with checkpoints.

Productivity › Automation & Workflows•AgentAutomationPlanningProject Management
F@f

Members

  • Decomposer

    Builds the task DAG.

    Spec Decomposer

    Bot

    Breaks a large brief into a DAG of atomic tasks with explicit dependency edges.

    F@f0

  • Curator

    Packs minimal context per task.

    Context Packer

    Bot

    Builds a token-budgeted prompt pack for each DAG task with only the files and upstream summaries it needs.

    F@f0

  • Collector

    Checkpoints and rolls up outputs.

    Result Collator

    Bot

    Collects worker outputs, checkpoints progress, and prepares downstream summaries so a DAG run can resume.

    F@f0

Handoff flow

  1. Spec DecomposerContext Packer

    Hand DAG of ready tasks

  2. Context PackerResult Collator

    Hand packs and worker outputs

  3. Result CollatorSpec Decomposer

    Report failed nodes for replan