harness-files/skills/rad-orchestration/SKILL.md
This skill can be invoked when the user starts talking about planning or executing a project. Use this skill to run a team of subagents through an automated SDLC process that helps them get work done safely and effectively.
npx skillsauth add MetalHexx/RadOrchestration rad-orchestrationInstall this skill globally with one command. Works with Claude Code, Cursor, and Windsurf.
3 of 9 scanners reported clean
Some scanners were skipped, did not run, or reported a non-clean status. Review each row below.
You are an orchestrator agent. Your job is to help the user with running projects through the
planning and execution subagent pipeline. Use this skill when you think the user is ready to plan
or execute a project from the ~/.radorc/projects/ directory. You will be responsible for running the pipeline, which includes signaling events, executing actions, and managing the overall flow of the project.
Read through all sections below before you get started on any project work.
Always read through the references/context.md document to understand the overall system and your role in it.
Before you can run the pipeline, you need to understand the system configuration and user preferences. This configuration is stored in ~/.radorc/orchestration.yml. Read this and commit it to memory. These values are important for the life of a project run.
The pipeline script (scripts/radorch.mjs) is the engine that drives the project forward. It listens for events, gives you action to execute, and manages the flow of the project. Familiarize yourself with how to run this script and how it interacts with the rest of the system. Read through the references/pipeline-guide.md document to understand the details of how the pipeline works, including the envelope structure, CLI commands, and error handling.
Review the documentation standards at references/document-conventions.md to understand the naming conventions, placement, and content of these documents. These documents are the audit trail of our projects.
I'd like you to avoid reading the Requirements, Master Plan, Phase Plan or Task Handoffs unless it's absolutely necessary to keep the project running smoothly or mandated by the pipeline prompt.
There is one exception to the document reading rule: for corrective mediation. So you'll read the responses and review reports generated by reviewer agents. When you receive a code_review_completed or phase_review_completed event with a verdict: changes_requested|rejected, read the relevant review documents and the references/corrective-playbook.md to learn how to mediate a potential correction.
radorch.mjs script to manage the flow of the project.
data.has_custom_instructions field.
rad-log-error skill if you run into any issues that you think are system errors or bugs.
state.json directly if you can avoid it.
~/.radorc/orchestration.yml and the pipeline results.
radorch.mjs script.schemas/orchestration-state-v5.schema.json — State file JSON Schemascripts/radorch.mjs CLI — Pipeline runtime and many other tools.~/.radorc/orchestration.yml — System configuration;~/.radorc/templates/ — The review-intensity tier process templates.testing
Use this skill to understand, navigate, and organize the work-graph — the live map of projects in the system, what state each is in, how they relate to one another, and where their worktrees are. Use at brainstorm time when continuing a series or referencing existing work — orient on live state before exploring files.
development
Source control operations — commit code, open a PR, create a worktree, or clean up a worktree. All inputs for commit and PR come from the spawn prompt; worktree operations are interactive and driven from the main session.
testing
Continue a project through the orchestration pipeline. Ensures the Orchestrator runs as the primary agent — not as a subagent — so it retains full control of agent sequencing. Use for local, background, or cloud-based execution.
tools
Set up a parallel git worktree for a project and launch orchestration execution in it. Use when asked to "run in parallel", "create a worktree", "execute in a worktree", or when launching a project in an isolated branch for parallel development. Handles git worktree creation, branch setup, source control initialization, and opening the worktree in VS Code, Copilot CLI, Claude Code, or a terminal.