liveBook · Manning

Implementation plan

As explained in Claude at Work

An implementation plan is a structured description of the work, decisions, and sequence required to build a software feature or system change.

An implementation plan describes how to build one task or vertical slice from a delivery plan, consistent with the overall strategy and design. It makes the proposed work visible for review before implementation begins. section 6.1 section 6.7

Planning in stages is not a commitment to a long, fixed delivery. Plans can be generated and revised quickly; smaller slices keep the work manageable and give the tech lead opportunities to steer before substantial code is written. section 6.1 section 6.7

Use a plan to prepare and control a slice:

  • Start with the strategy, delivery plan, and design documents. Ask the agent to reconcile them and identify gaps before proposing the build. section 6.7 section 6.8
  • Review the proposed approach, resolved gaps, and how the slice runs. Discuss open decisions and revise the plan before authorizing execution. section 6.0 section 6.8
  • Set explicit review gates. For a BDD slice, the plan can stop after producing the Gherkin feature so a person can review it before work continues. section 6.8

Plan the first slice

In Claude Code plan mode, a focused request can refer to existing constraints: “docs/plans/ch06/02-messaging-core-delivery-plan.md — plan the first slice. We will implement from this, so confirm the delivery plan against the design and close any gaps.” Review the result and resolve discrepancies before saying, “Let’s execute.” section 6.8

For several slices, parallel planning needs coordination. Review dependencies and open decisions, define each subagent’s role, and give a central agent responsibility for checking consistency. Decide whether unresolved questions require a human stop or may be decided and reported back. section 6.10

Listing 6.6 Part of the prompt Claude wrote for me to send to the next Claude
Do this yourself, before dispatching anything. Write
`docs/plans/ch06/04-slice-contracts.md`.

Six planners working blind will close the same gap two different ways
wherever slices touch. This document is the answer, and it is the reason
the fan-out is safe. It is not a summary of the delivery plan — it is a
set of **rulings**, each naming the one slice that owns the thing.

A shared contracts document assigns ownership where slices touch, helping parallel planners avoid incompatible decisions. section 6.10

How Claude at Work covers it

How deep it goes
Works through plan inputs, gap reconciliation, human review, execution gates, and coordination of parallel planners.
Examples
Uses a first-slice messaging plan, a Gherkin review gate, and parallel plans coordinated through shared slice contracts.
What it assumes
Assumes existing strategy, delivery, and design documents that supply constraints for the slice plan.
What you can do after
Operate Claude Code’s plan mode to review a slice plan, resolve gaps, and authorize execution with a review gate.
How it connects
Connects strategy and delivery planning with vertical slicing and behavior-driven development.

Related concepts

This explainer was generated by Manning's knowledge graph from the text of Claude at Work. The section links point to the parts of the book it draws on.