Starlight Autoplan

Starlight Autoplan

Purpose

Turn a rough proposal into a bounded plan whose assumptions, risks, authority limits, and verification gates are visible before implementation begins.

When it fires

• The user asks to review, harden, or stress-test a plan.
• The work crosses product, data, security, design, runtime, or deployment boundaries.
• A wrong decision would be expensive or difficult to reverse.

Do not invoke the full review for a trivial, reversible edit unless the user asks for it.

Inputs

• Desired user outcome and explicit non-goals.
• Current repository or system evidence.
• Time, cost, compatibility, privacy, and authority constraints.
• Known failure history and required verification commands.

Workflow

1. Restate the outcome, success criteria, non-goals, and unresolved assumptions.
2. Select review depth proportional to reversibility and impact.
3. Run four lenses: product value, engineering integrity, design quality, and developer/operator experience.
4. Give every finding an evidence reference, confidence level, and consequence.
5. Reconcile contradictions; never silently average incompatible recommendations.
6. Produce the smallest coherent execution sequence with rollback and promotion gates.
7. Stop for human direction before destructive, costly, public, or credential-bearing actions.

When the same agent performs every lens, disclose that the review was sequential rather than independent.

Output contract

Return:

• decision: build, revise, defer, or stop.
• scope: user-visible outcome, non-goals, and affected surfaces.
• assumptions: verified, inferred, and unresolved items kept separate.
• reviews: findings from each lens with severity and evidence.
• execution_plan: ordered, independently verifiable steps.
• verification: automated, manual, security, accessibility, and deployment gates as applicable.
• artifacts: plan, decision, rollback, and verification records with stable references.
• failure_conditions: conditions that force revise, defer, or stop.
• verification_receipts: one receipt per applicable gate with gate, result, evidence, and owner.
• rollback: safe recovery path and ownership.
• human_gates: decisions the agent may not make.

Return decision: build only when every gate applicable before implementation has a passing verification receipt. Record later-stage gates as pending; when their decision point arrives, a failed or missing applicable receipt blocks promotion and requires revise, defer, or stop.

Tools & MCP

Use repository search, tests, type/lint gates, preview deployments, and an approved browser connector when relevant. Do not enable connectors, spend money, publish, merge, or mutate external systems merely because a plan mentions them.

Quality bar

• No claim that AI makes execution free or removes operational ownership.
• No speculative expansion without a stated user outcome.
• Evidence and inference are distinguishable.
• Every high-risk step has a verifier and rollback.
• The plan can be handed to another operator without hidden context.

Example

Input: “Add a self-service agent marketplace with portable plugins.”

Good output: a revise decision that separates catalog, package, installer, evaluation, and promotion contracts; identifies marketplace approval and credential handling as human gates; and gives exact API, schema, preview, security, and rollback checks.

Built on SIP — Starlight Intelligence Protocol Vertical: starlight-agent-skills · portable capability layer