- Home
- Skills
- Productivity
- Plan Master: Systematic Planning
More screenshots
Works with the AI Agents you already use
Plan Master: Systematic Planning
Plan complex work before executing: clarify the outcome in checkable terms, map constraints and unknowns, sequence verifiable slices, premortem the risks, and validate before building.
$5
Plan Master: Systematic Planning
Example session with this skill installed
We're a four-person SaaS team. Our project tool needs team workspaces with shared billing — multiple users under one account, one invoice per workspace. Six-week deadline, two engineers, a monolith, and one Postgres database everything shares. Plan it before we write code.
- Read your context and instructions
- Compiled the plan master
GOAL — Workspace admins can invite users into a workspace, assign roles, and receive one consolidated invoice per workspace per month. Success: a pilot customer runs their whole team (8 users) on one workspace for a full billing cycle; individual accounts keep working unchanged.
NON-GOALS — No per-seat proration in v1 (seats billed at cycle start). No workspace switching for users in multiple workspaces (v2). No migration of existing invoices.
DECISIONS — Workspace becomes a tenancy column on users/projects/billing rather than a separate service: rejected a billing microservice (two engineers, six weeks, monolith is fine). Billing stays with the current payment provider's subscription-per-workspace model: rejected one subscription per user (invoice fragmentation is the complaint we're fixing).
UNKNOWNS — Provider's support for quantity changes mid-cycle → half-day spike in week 1 against their sandbox, before schema work depends on it. How existing solo accounts map to workspaces → assumption: every account becomes a one-person workspace; validated by a data audit query in week 1.
SEQUENCE — 1. Spike: provider sandbox, quantity change + consolidated invoice → checkpoint: a test invoice shows 8 seats, one line total.
2. Schema: workspace tenancy column, backfill solo accounts → checkpoint: backfill runs on a production copy, zero orphaned rows.
3. Invites + roles end-to-end for one workspace → checkpoint: invited user signs in, sees only that workspace's projects.
4. Billing switch behind a per-workspace flag → checkpoint: flagged workspace invoiced by the new path; unflagged accounts unchanged.
5. Pilot rollout → checkpoint: one full cycle, invoice matches manual count.
RISKS — Backfill corrupts existing billing rows → mitigation: additive columns only, old fields untouched until pilot completes. Provider spike fails → fallback: manual seat-count sync job, uglier but shippable. ROLLBACK — flag off per workspace restores per-user billing; schema changes are additive, so nothing needs reversing.
Connects securely to your tools. The creator never sees your data.
About this skill
A complete planning method for work where being wrong is expensive — multi-step features, migrations, anything spanning sessions or people. It starts from a simple definition: a plan is decisions made in advance plus a sequence with checkpoints, and everything else is decoration. The method walks six moves. Clarify the outcome first — the problem in one paragraph, success criteria as checkable statements ("a user can do X and observe Y," not feature names), and explicit non-goals, because writing down what's out of scope is the cheapest scope control there is. Then map the territory: existing pieces to reuse, unknowns listed honestly with a resolution path for each (a time-boxed spike, a question, or a labeled assumption), key decisions with the rejected options and reasons, and dependencies named before they kill a schedule. Sequencing favors vertical slices over horizontal layers, puts the scariest unknown before the work that depends on it, and attaches a checkpoint to every step — a step without one is a hope. A premortem assumes the plan already failed and works backward to concrete causes, converting each into a prevention or a mitigation. The whole plan fits a one-screen template: goal, non-goals, decisions, unknowns, sequence, risks, rollback. A decision table settles the plan-vs-execute question up front (small, reversible, one-file work should just be done), a validation checklist and a 14-point quality scorecard gate the plan before it is presented, and the worked example shows the full method producing a validated plan for adding SSO to a B2B SaaS app — pilot customer, SAML spike first, break-glass path before enforcement, flag-gated rollback.
Find more skills like this on the Claude Skills Marketplace, the Codex Skills Marketplace and the Gemini Skills Marketplace, or browse Claude Skills for Productivity.
- The plan is only as honest as its inputs — an unknown nobody mentions can't be mapped, which is why validation with the user is a step, not a courtesy.
- It deliberately produces a one-screen plan, not a full specification; work that needs a spec still needs one after the plan is validated.
- Estimates, where given, are sequencing aids, not commitments — the method verifies with checkpoints instead of trusting durations.
How to use
- 1
Connect your agent to the Agensi MCP
Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup
- 2
Sign in to Agensi
Your agent shows a sign-in link. Sign in and click Allow.
- 3
Ask for what you need
Your agent loads this skill and uses it on your task. It always asks before buying a skill.
Reviews
No reviews yet
Be one of the first to try it. Every listed skill goes through the trust checks below.
Security scanned
Checked by our 8-point scan before listing
Fresh listing
Recently published to Agensi
Money-back guarantee
Doesn't do what its page promises? You get your money back.
Trust & safety
Security scan
Security scan passed 24 days ago
Permissions
No special permissions declared or detected