agentsclimarketplace

Planning with artifacts

Skill IrfanSadiqRahat/constellation/methodology/planning-with-artifacts

Break work into 2–5 min tasks, each ending in a typed artifact handoff. Activates after ScopedBrief.From its SKILL.md

Install
npx -y skills add IrfanSadiqRahat/constellation --skill planning-with-artifacts

Assembled from the repository path, not quoted from the project. Check it against their README if it does not work.

2 things to look at

  • no licenseNo license file was found in the repository. Code published without one is not open source by default, so using it at work is a question for whoever answers licensing questions where you are.
  • 1 stars1 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.

SKILL.md

2.3 KB, 566 tokens by cl100k_base, as published. Nobody here has run it

planning-with-artifacts

The deliverable: Plan

goal: <from ScopedBrief>
phases:
  - id: P1
    role: <which agent runs this>
    consumes: <input artifact type>
    produces: <output artifact type>
    tasks:
      - id: P1-T1
        title: <2–5 min unit>
        verification: <observable, single boolean>
        files: [<exact path>]
        notes: <constraints, gotchas>
risk_register: [<risk + mitigation + owner>]
rollback_points: <which phases support rollback, how>
done_criteria: <observable, in numbers>

Operating principles

  1. Tasks are atomic units of execution. 2–5 minutes of agent time. Bigger = split.
  2. Every task ends in a verification. Boolean. Not "looks good". Numeric or shape-checked.
  3. Files listed by exact path. "Update the auth module" is not a task. "Modify src/auth/login.ts:42 to reject empty tokens" is.
  4. Phases handoff via typed artifacts. P1 produces BackendArchitecture. P2 consumes it. No tribal context transfer.
  5. Risk register has owners. Risks without owners are not risks, they are wishes.
  6. Rollback points named. If P3 is destructive, P2 has to be revisitable.
  7. Done is numeric. "Acceptance ≥ 95% on eval set X" not "feels right".

Role-aware adaptation

The Plan's phases are sequenced by the team formation in use. If the user said team saas-mvp, the Plan inherits the SaaS-MVP pipeline structure:

P1: discover    role=product-strategist+prd-writer   →PRD
P2: design      role=design-lead                     →DesignBrief
P3: build       role=frontend-architect+backend-architect+db-architect  →MVP
P4: monetize    role=billing-architect               →MonetizedMVP
P5: harden      role=sre-lead+privacy-architect      →ProductionMVP
P6: launch      role=growth-lead                     →LaunchPlan

Each phase contains the role-specific task taxonomy.

Why this skill is opinionated

  • Typed artifact handoff between phases — no tribal context loss
  • Role-aware task taxonomy — frontend tasks look different from data tasks
  • Rollback points as a first-class concept
  • Risk register with owners, not afterthoughts

What ships with it

Read from the repository

Just SKILL.md. No reference files, no scripts.

Keep looking

Skills are one crate of 326,790. Ordering is by how many stacks a row turns up in, so the top of any crate is what has actually been picked rather than what has the most stars.