---
name: multi-agent-local
language: en
description: "Full pipeline in local mode  -  no worktree, runs directly on the current branch. Use when the full pipeline should run on the current branch without creating a worktree."
user-invocable: true
---

# multi-agent local  -  Full Pipeline, Local Branch

Runs the full 8-phase pipeline (normal mode  -  including the Plan Approval Gate + parallel review + triage) on the current branch **without creating a worktree**. The dedicated equivalent of the `multi-agent "task" --local` flag form.

## When to use it

- You do not want an extra open worktree  -  if an editor/IDE is open in the same folder, it is not disrupted
- You are already on the correct branch and just want to apply the pipeline discipline
- Small project / prototype  -  worktree overhead is unnecessary

## When NOT to use it

- Multiple tasks in parallel at the same time  -  worktree isolation is lost
- Multi-repo task  -  `--local` locks to a single repo

## Pipeline

The same 8 phases as the normal pipeline: Init → Analysis → Planning → Dev → Review → Test → Commit → Report. Only Phase 0 Step 8 (workspace creation) continues on the current branch instead of a worktree; the `state.projects[*].worktreePath` field stays `null`.

## Delegation

Delegates to the orchestrator skill (`multi-agent/SKILL.md`) with the `--local` flag. The Phase 0 Step 6 worktree step is skipped; all other phase contracts apply as is (Plan Approval Gate, three-reviewer review, triage, secret scan, PR creation).

## Examples

```bash
multi-agent-local "PROJ-12345"                                 # Jira
multi-agent-local "#42"                                        # GitHub issue
multi-agent-local "LoginView dark mode fix"                    # Free-text
```

## Required: outward-facing payload contracts

Before writing anything outward-facing  -  PR body, Jira comment, Confluence page, closing report  -  load `$HOME/.claude/multi-agent-refs/payload-contracts.md`. It names the canonical section set for each payload, the markup dialect per surface (PR body is Markdown, Jira is wiki markup  -  mixing them is a defect), and the token/duration numbers the closing report must carry. Improvising a payload shape from memory is the most common failure of the short modes.