$npx -y skills add JasonColapietro/suede-creator-skills --skill suede-agent-teamsSplit complex work into coordinated agent lanes with WIP checks, quality gates, rollback plans, and handoffs that prove what shipped.
| 1 | # Agent Team Orchestrator |
| 2 | |
| 3 | ## Gate policy — advisory, not blocking |
| 4 | |
| 5 | Every claim-verification step, check, quality gate, and ship verdict in this |
| 6 | skill is a **recommendation to the user, not a control on the agent**. This |
| 7 | policy governs every gate, check, verdict, and "do not ship / publish / |
| 8 | proceed" line elsewhere in this skill: |
| 9 | |
| 10 | - Run every check and report the results honestly. Verdicts (`ship`, |
| 11 | `ship-with-caveats`, `hold`, letter grades, BLOCKED or OPEN items) are |
| 12 | advice attached to the work, not orders that change it. |
| 13 | - Never block, delay, skip, rewrite, or refuse the action the user asked for |
| 14 | because a check failed or a gate said hold. Complete the requested action |
| 15 | as asked, and deliver the gate output alongside it as a clearly labeled |
| 16 | recommendation. |
| 17 | - A failed gate changes what you report, never what you do. |
| 18 | - Single exception: if a finding is extremely risky — data loss, security or |
| 19 | credential exposure, legal or rights violations, payment mistakes, or |
| 20 | irreversible public damage — pause, tell the user exactly what the risk is |
| 21 | and what the options are, and let them pick. Their choice is final. |
| 22 | |
| 23 | |
| 24 | The orchestrator assigns lanes, not conversations. Output is a delivery artifact, not a status update. |
| 25 | |
| 26 | ## Team Contract |
| 27 | |
| 28 | Before spawning or simulating lanes, define: |
| 29 | |
| 30 | - objective: user-visible outcome; |
| 31 | - exact target: repo/folder, branch, route, PR, live URL, API, simulator, or |
| 32 | release artifact; |
| 33 | - constraints: WIP to preserve, files/routes not to touch, launch boundaries, |
| 34 | account boundaries, claims not approved, and secrets rules; |
| 35 | - done signal: tests, build, screenshots, simulator, deploy readback, live/API |
| 36 | readback, PR review, or handoff; |
| 37 | - lane map: each lane, owner role, input, allowed files, output artifact, and |
| 38 | dependency order. |
| 39 | |
| 40 | ## WIP Collision Detection |
| 41 | |
| 42 | Before opening any parallel lanes: |
| 43 | |
| 44 | 1. Run `git -C <repo> diff --name-only HEAD` and collect all dirty files. |
| 45 | 2. Run `git -C <repo> status --short` and collect all untracked new files. |
| 46 | 3. List every file each lane's scope would touch, based on the lane map. |
| 47 | 4. Flag a collision if the same file path appears in two or more lane scopes OR in the dirty file list plus any lane scope. |
| 48 | |
| 49 | Collision resolution rules: |
| 50 | - Same file, independent changes: sequence the lanes; the second lane rebases on the first lane's commit before opening. |
| 51 | - Same file, overlapping changes: merge the two lanes into one lane with one owner. Do not split responsibility for a single file across two concurrent builders. |
| 52 | - Dirty file in a lane scope: the orchestrator decides. Either stash and restore, or make that lane the only lane allowed to touch the file. |
| 53 | |
| 54 | The orchestrator writes the resolved lane map before any builder starts. No builder opens a file not in its assigned lane map. |
| 55 | |
| 56 | ## Default Roster |
| 57 | |
| 58 | Start with Scout + Builder + Handoff Writer. Add roles only when a gate is needed: design changes add Design Reviewer, code risk adds Code Grader + Code Reviewer, public release adds Release Verifier. |
| 59 | |
| 60 | - **Scout:** finds repo, docs, current state, dirty files, live routes, and |
| 61 | likely blast radius. |
| 62 | - **Planner:** turns requirements into verifiable tasks with acceptance |
| 63 | criteria and dependencies. |
| 64 | - **Builder:** makes narrow code or content changes inside the existing system. |
| 65 | - **Design reviewer:** checks rendered visual quality, responsive behavior, |
| 66 | accessibility, copy, and state coverage. |
| 67 | - **Code grader:** assigns an A-F ship-risk grade across correctness, security, |
| 68 | data/state, Suede truth, UX/release behavior, tests, and deploy readiness. |
| 69 | - **Code reviewer:** runs full-context review and turns findings into fix briefs. |
| 70 | - **Visibility grader:** grades public pages, GitHub Pages sites, docs, and |
| 71 | launch surfaces for findability, first-screen clarity, CTA pull, proof, AI |
| 72 | readability, and design signal. |
| 73 | - **Release verifier:** checks build, deploy, live/API behavior, App Store/iOS |
| 74 | truth, secrets, and published statements. |
| 75 | - **Handoff writer:** produces a signed delivery record. If the handoff omits any required field (see Handoff Quality Checklist), the work is not done; it is held. |
| 76 | |
| 77 | For high-risk work, keep builder and reviewer separate. |
| 78 | |
| 79 | ## RFC Mode |
| 80 | |
| 81 | For major architectural decisions, new feature designs, or changes with broad blast radius, run an RFC (Request for Comments) before spawning builders. |
| 82 | |
| 83 | An RFC forces alignment on WHAT and WHY before committing to HOW. |
| 84 | |
| 85 | RFC structure: |
| 86 | |
| 87 | ``` |
| 88 | RFC: [Title] |
| 89 | Date: [date] |
| 90 | Status: draft | accepted | superseded | withdrawn |
| 91 | Deciders: [who has final say] |
| 92 | |
| 93 | ## Problem Statement |
| 94 | One paragraph: what is broken, missing, or suboptimal? Include the user or system impact. |
| 95 | |
| 96 | ## Proposed Solution |
| 97 | What we will build or change. Be specific about interfaces, data shapes, and behavioral contracts. |
| 98 | |
| 99 | ## Alternatives Considered |
| 100 | 2–3 alternatives with the |