Recommend and optionally run the next factory stage for this repo
Scanned 9/2/2026
Install to Claude Code
npx -y skills add watt-mind/factory --skill factory-next --agent claude-codeInstalls into .claude/skills of the current project.
Are you the author of Factory Next?
Add the live security badge to your README — it updates automatically with every re-scan.
[](https://www.skillsdirectory.com/skills/watt-mind-factory-next)More formats (shields.io, HTML) on the badges page.
---
name: factory-next
description: Recommend and optionally run the next factory stage for this repo
---
# factory-next
The user's accompanying request is this workflow's argument string. Wherever these instructions refer to `$ARGUMENTS`, interpret it as that request.
Pick **one** next step in the factory loop for the current repository — diagnose by default, run only when asked.
The mechanical routing lives in `factory next`, which reads the same queue snapshot as `factory queue`. Do not re-derive gates from Linear yourself.
## 1. Resolve repo
Default: match the current working directory against `config/repos.yaml`. `$ARGUMENTS` may name a repo (`bj29`) or pass flags (`--apply`, `--orchestrated`, `--include-sweep`, `--harness pi`).
## 2. Diagnose
From the factory checkout (or any cwd with `factory` on PATH):
```bash
factory state --repo <repo>
factory next --repo <repo> $ARGUMENTS
```
Show **session history** from `factory state` (what ran earlier in this harness window), then **constraint**, **recommend**, **because**, and any **also** alternates from `factory next`. Both record events to `~/.factory/state/events.jsonl` — `recommend` on every next run, `start` on `--apply`.
If `budgetBlocked` is set, stop — no new agent stages today.
## 3. Apply (only when `$ARGUMENTS` contains `--apply`)
**One bounded stage, then stop.** Do not loop until idle — that is the supervisor's job, not `/factory-next`.
- If the recommendation is a slash command (`/factory-merge`, `/factory-triage`, `/factory-work`, …), run **that command's skill/body** in this session — one pass, default caps (`5` triage, `3 at a time` work, `10` unblock).
- If `--orchestrated` is present and the recommendation is `tick` or `reconcile`, run the printed `exec` line instead.
- If the recommendation is `(wait)` with no command, report why and do not spawn work.
Alternatively, delegate execution to the script:
```bash
factory next --repo <repo> --apply [--orchestrated] [--harness <harness>] $OTHER_FLAGS
```
Use the script path when running headless via `run-agent.sh`; use the slash command path when you are already in an interactive harness.
## 4. Report
State what ran (or what was recommended), the queue constraint in one line, and anything waiting on a human (blocked holds — suggest `factory digest --repo <repo>`).
Record the outcome:
```bash
factory state record --type complete --repo <repo> --command <stage> --summary "<one line>"
```
**Session friction:** when `FACTORY_RUN_ID` is unset, scan the session per `/factory-friction` and note items filed or **none observed**. If any were filed:
```bash
factory state record --type friction --repo <repo> --issues "OPS-123"
```
## Priority (for transparency — the script encodes this)
1. Held tickets with replies → triage
2. Open mergeable PRs on GitHub → merge
3. Free dispatch slots + startable tickets → work (or `tick.mjs` when orchestrated)
4. Triage / Todo-not-ready backlog → triage
5. Orphaned claims → reconcile (alternate)
6. Holds without answers → digest / unblock (alternate)
7. Main loop idle + `--include-sweep` → sweep
Retro, ship, audit, and friction capture are **not** auto-selected — run those on cadence or explicitly.
Is this your skill, or is something wrong with this listing? Request removal or report an issue. Author removals are honored within 72 hours.
No comments yet. Be the first to comment!