Operant Studio
OPER-6

extract(runners): move agent-*.yml workflows from Portico to Ledger as reusable workflows

Band
LARGE

Body

## Problem
Portico still owns 8 fat local Runner workflows totaling ~1500+ lines with zero external references:
- agent-auto-ready.yml (308 lines)
- agent-automerge.yml (88)
- agent-coderabbit-trigger.yml
- agent-dispatch.yml (428)
- agent-mark-done.yml
- agent-pr-sweeper.yml
- agent-rebase.yml
- agent-size-override.yml

These are studio-layer primitives. Every future Operant app (Piper DX, Roster, Pillars) will need them. Today they are copy-paste-fork bait.

## Scope
1. Add `.github/workflows/runners/agent-*.yml` in operant-studio with `on: workflow_call:` and typed inputs (repo, task_id, sha, etc.)
2. Convert each Portico workflow to a one-line shim: `jobs: run: uses: Operant-Health/operant-studio/.github/workflows/runners/agent-dispatch.yml@main` with `secrets: inherit`
3. Test agent-automerge (smallest) end-to-end before converting the rest
4. Piper DX / Roster / Pillars / new products then just `uses:` the same runners

## Depends
- Nothing.

## Blocks
- Any future "add Piper DX to autonomous dev" work.

## Verification
- Kick a test task through agent-dispatch via the shim.
- PR opens.
- agent-automerge runs and reports.
- Logs in Actions UI show reusable-workflow ID.

Attachments

Loading attachments…

Comments

Loading comments…