Skip to content
Get packs

Graph Pipeline Runner

Runs repeatable multi-step pipelines from a JSON plan, resuming after crashes without repeating finished steps.

Source Download skill
Repository
Yeachan-Heo/oh-my-claudecode
Stars
40,512
Licence
MIT
Files
1 file

What it does

This skill lets the agent turn a multi-step task into a fixed JSON workflow (a directed acyclic graph of command and agent steps) and run it through the oh-my-claudecode command line tool. Each step's progress is written to a journal, so if the run is interrupted, rerunning the same command resumes from where it stopped instead of starting over. Steps that need a human yes/no can pause for approval, but only when run interactively.

Great for

  • Developers who want the same multi-step build, test, or check sequence run reliably and repeatably
  • Teams that need an auditable record of which pipeline steps ran and their outcomes
  • Anyone running long pipelines that might be interrupted and need safe resumption

Skip it if

  • Exploratory or one-off tasks, since the graph is fixed and cannot re-plan mid-run; ordinary conversation is more suitable
  • Pipelines that run untrusted or unreviewed shell commands, because command steps have full authority in the working directory and are not sandboxed
  • Workflows needing agent steps that edit files or call external services, since built-in agent steps are read-only and such work requires writing a custom executor

Try asking

  • “Build, then test, then ask me before deploying this project”
  • “Create a pipeline that lints, runs unit tests, and runs an agent review of the diff”
  • “Resume the nightly test pipeline I started before my machine restarted”
Step-by-step guide: how to use Graph Pipeline RunnerInstall, first run, examples, tips and FAQs

Install in 3 steps

  1. 1

    Download the skill

    The ZIP contains the 1 original files, the MIT licence and an attribution note.

    graph.zip
  2. 2

    Add it to your agent

    Claude Code (all projects)

    mkdir -p ~/.claude/skills && unzip graph.zip -d ~/.claude/skills/

    Claude Code / Agent SDK (this project only)

    mkdir -p .claude/skills && unzip graph.zip -d .claude/skills/

    Claude apps (web/desktop): Settings → Capabilities → Skills → Upload skill, and choose the ZIP.

  3. 3

    Ask for the task

    No special command needed — the agent reads the skill's description and loads it when your request matches. Try one of the prompts above.

Get better results

  • Read and review the generated descriptor JSON before running it, since it is executable content and command nodes run shell lines.
  • Use a unique run_id per pipeline; reusing one resumes the old run rather than starting fresh.
  • For approval gates, run the command with the ! prefix in an interactive session so y/n input works.

What's inside

  • SKILL.md

Redistributed unmodified from Yeachan-Heo/oh-my-claudecode under the MIT licence. The assessment was drafted with AI from the skill's files on 2026-10-29. Skills can instruct your agent to run code — read SKILL.md and any scripts before installing.