Stacks / 7 workflow stages · v3.1.0 · reviewed 2026-09-05

Use AI on an Existing Codebase

Give the coding agent the right repository context, current docs, architecture references, and browser-level verification.

Best for

Developers using AI to make meaningful changes in an existing repository without letting the agent operate from shallow context.

Outcome

A coding workflow that understands the repository, retrieves current documentation, makes changes, reviews them, and verifies behavior.

Not for

A greenfield prototype where speed matters more than understanding an established architecture.

Operator guidance

Best whenThe repository has tests and review gates the coding agent can run.

Avoid whenThe task lacks a reproducible issue, ownership boundaries, or a safe verification path.

CompatibilityExpose repository conventions, current dependency docs, and the same checks used in CI; isolate agent changes in branches or worktrees.

Use this as a decision framework, not a mandatory shopping list. Swap or skip layers based on your existing stack, constraints, and risk profile.

01

1. Choose the coding agent

Start with a capable terminal-native coding workflow that can inspect and modify a real repository.

Alternatives
  • Cursor
    Recommendation details

    Why: Consider Cursor as an alternative for 1. Choose the coding agent; compare its documented capabilities and operating constraints with the primary recommendation.

    Best when: Popular AI-native code editor. Useful for comparing agent UX, workflows and rules.

    Avoid when: Avoid when the current official source does not support your required deployment, data-handling, or integration constraints.

    Compatibility: Validate interfaces, data formats, and operational handoffs with the adjacent stack stages.

    Prerequisites:

    • Review the current official source and implementation requirements.

    Evidence: Cursor official source (2026-09-05)

    Verified 2026-09-05 · Medium confidence

  • OpenAI Codex
    Recommendation details

    Why: Consider OpenAI Codex as an alternative for 1. Choose the coding agent; compare its documented capabilities and operating constraints with the primary recommendation.

    Best when: Open-source terminal coding agent; useful for comparing agent workflows and implementation patterns.

    Avoid when: Avoid when the current official source does not support your required deployment, data-handling, or integration constraints.

    Compatibility: Validate interfaces, data formats, and operational handoffs with the adjacent stack stages.

    Prerequisites:

    • Review the current official source and implementation requirements.

    Evidence: OpenAI Codex official source (2026-09-05)

    Verified 2026-09-05 · Medium confidence

  • Aider
    Recommendation details

    Why: Consider Aider as an alternative for 1. Choose the coding agent; compare its documented capabilities and operating constraints with the primary recommendation.

    Best when: Terminal pair-programming tool with mature Git-oriented workflows.

    Avoid when: Avoid when the current official source does not support your required deployment, data-handling, or integration constraints.

    Compatibility: Validate interfaces, data formats, and operational handoffs with the adjacent stack stages.

    Prerequisites:

    • Review the current official source and implementation requirements.

    Evidence: Aider official source (2026-09-05)

    Verified 2026-09-05 · Medium confidence

Recommendation details

Best when: Official Claude Code documentation.

Avoid when: Avoid when the current official source does not support your required deployment, data-handling, or integration constraints.

Compatibility: Validate interfaces, data formats, and operational handoffs with the adjacent stack stages.

Prerequisites:

  • Review the current official source and implementation requirements.

Evidence: Claude Code Docs official source (2026-09-05)

Verified 2026-09-05 · Medium confidence

02

2. Package repository context

Create a compact representation of a repository when you need to reason across files or share codebase context cleanly.

Alternatives
  • Gitingest
    Recommendation details

    Why: Consider Gitingest as an alternative for 2. Package repository context; compare its documented capabilities and operating constraints with the primary recommendation.

    Best when: Turn a GitHub repository into a prompt-friendly text digest in seconds.

    Avoid when: Avoid when the current official source does not support your required deployment, data-handling, or integration constraints.

    Compatibility: Validate interfaces, data formats, and operational handoffs with the adjacent stack stages.

    Prerequisites:

    • Review the current official source and implementation requirements.

    Evidence: Gitingest official source (2026-09-05)

    Verified 2026-09-05 · Medium confidence

Recommendation details

Best when: Pack an entire repo into AI-friendly context, with token counting, compression and secret scanning.

Avoid when: Avoid when the current official source does not support your required deployment, data-handling, or integration constraints.

Compatibility: Validate interfaces, data formats, and operational handoffs with the adjacent stack stages.

Prerequisites:

  • Review the current official source and implementation requirements.

Evidence: Repomix official source (2026-09-05)

Verified 2026-09-05 · Medium confidence

03

3. Understand unfamiliar code

Use generated repository documentation and structure to orient the agent before large or architectural changes.

Alternatives
  • grep.app
    Recommendation details

    Why: Consider grep.app as an alternative for 3. Understand unfamiliar code; compare its documented capabilities and operating constraints with the primary recommendation.

    Best when: Search real code across public GitHub repositories to find how libraries and patterns are actually used.

    Avoid when: Avoid when the current official source does not support your required deployment, data-handling, or integration constraints.

    Compatibility: Validate interfaces, data formats, and operational handoffs with the adjacent stack stages.

    Prerequisites:

    • Review the current official source and implementation requirements.

    Evidence: grep.app official source (2026-09-05)

    Verified 2026-09-05 · Medium confidence

Recommendation details

Best when: Instant conversational documentation for public GitHub repositories. Excellent for unfamiliar codebases.

Avoid when: Avoid when the current official source does not support your required deployment, data-handling, or integration constraints.

Compatibility: Validate interfaces, data formats, and operational handoffs with the adjacent stack stages.

Prerequisites:

  • Review the current official source and implementation requirements.

Evidence: DeepWiki official source (2026-09-05)

Verified 2026-09-05 · Medium confidence

04

4. Pull current library docs

Give the agent current implementation documentation instead of relying on model memory for fast-changing libraries.

Alternatives
  • Postman Public API Network
    Recommendation details

    Why: Consider Postman Public API Network as an alternative for 4. Pull current library docs; compare its documented capabilities and operating constraints with the primary recommendation.

    Best when: Large collection of public API workspaces and documentation Claude can inspect for integrations.

    Avoid when: Avoid when the current official source does not support your required deployment, data-handling, or integration constraints.

    Compatibility: Validate interfaces, data formats, and operational handoffs with the adjacent stack stages.

    Prerequisites:

    • Review the current official source and implementation requirements.

    Evidence: Postman Public API Network official source (2026-09-05)

    Verified 2026-09-05 · Medium confidence

Recommendation details

Best when: Fresh library documentation for coding agents.

Avoid when: Avoid when the current official source does not support your required deployment, data-handling, or integration constraints.

Compatibility: Validate interfaces, data formats, and operational handoffs with the adjacent stack stages.

Prerequisites:

  • Review the current official source and implementation requirements.

Evidence: Context7 official source (2026-09-05)

Verified 2026-09-05 · Medium confidence

05

5. Follow agent coding practices

Use explicit patterns for context, delegation, verification, and long-running coding work rather than prompting ad hoc.

Alternatives
  • Cursor Agent Best Practices
    Recommendation details

    Why: Consider Cursor Agent Best Practices as an alternative for 5. Follow agent coding practices; compare its documented capabilities and operating constraints with the primary recommendation.

    Best when: Strong practical guidance for working with coding agents; many patterns transfer directly to Claude Code.

    Avoid when: Avoid when the current official source does not support your required deployment, data-handling, or integration constraints.

    Compatibility: Validate interfaces, data formats, and operational handoffs with the adjacent stack stages.

    Prerequisites:

    • Review the current official source and implementation requirements.

    Evidence: Cursor Agent Best Practices official source (2026-09-05)

    Verified 2026-09-05 · Medium confidence

Recommendation details

Best when: Anthropic's practical coding-agent workflow guidance.

Avoid when: Avoid when the current official source does not support your required deployment, data-handling, or integration constraints.

Compatibility: Validate interfaces, data formats, and operational handoffs with the adjacent stack stages.

Prerequisites:

  • Review the current official source and implementation requirements.

Evidence: Claude Code Best Practices official source (2026-09-05)

Verified 2026-09-05 · Medium confidence

06

6. Review the change

Add a deliberate review pass so generated changes are challenged before they are accepted.

Alternatives
  • Refactoring.Guru
    Recommendation details

    Why: Consider Refactoring.Guru as an alternative for 6. Review the change; compare its documented capabilities and operating constraints with the primary recommendation.

    Best when: Exceptionally clear reference for design patterns, SOLID and refactoring techniques.

    Avoid when: Avoid when the current official source does not support your required deployment, data-handling, or integration constraints.

    Compatibility: Validate interfaces, data formats, and operational handoffs with the adjacent stack stages.

    Prerequisites:

    • Review the current official source and implementation requirements.

    Evidence: Refactoring.Guru official source (2026-09-05)

    Verified 2026-09-05 · Medium confidence

  • Patterns.dev
    Recommendation details

    Why: Consider Patterns.dev as an alternative for 6. Review the change; compare its documented capabilities and operating constraints with the primary recommendation.

    Best when: Modern JavaScript, React and web-performance patterns with concrete examples.

    Avoid when: Avoid when the current official source does not support your required deployment, data-handling, or integration constraints.

    Compatibility: Validate interfaces, data formats, and operational handoffs with the adjacent stack stages.

    Prerequisites:

    • Review the current official source and implementation requirements.

    Evidence: Patterns.dev official source (2026-09-05)

    Verified 2026-09-05 · Medium confidence

PluginCode & Build
Recommendation details

Best when: Anthropic-verified multi-agent pull-request review with confidence filtering.

Avoid when: Avoid when the current official source does not support your required deployment, data-handling, or integration constraints.

Compatibility: Validate interfaces, data formats, and operational handoffs with the adjacent stack stages.

Prerequisites:

  • Review the current official source and implementation requirements.

Evidence: Code Review Plugin official source (2026-09-05)

Verified 2026-09-05 · Medium confidence

07

7. Verify in the browser

Test real user flows and rendered behavior instead of treating a successful build as proof the feature works.

Alternatives
  • Chrome DevTools Plugin
    Recommendation details

    Why: Consider Chrome DevTools Plugin as an alternative for 7. Verify in the browser; compare its documented capabilities and operating constraints with the primary recommendation.

    Best when: Give Claude DevTools-level access to performance traces, network requests and live browser debugging.

    Avoid when: Avoid when the current official source does not support your required deployment, data-handling, or integration constraints.

    Compatibility: Validate interfaces, data formats, and operational handoffs with the adjacent stack stages.

    Prerequisites:

    • Review the current official source and implementation requirements.

    Evidence: Chrome DevTools Plugin official source (2026-09-05)

    Verified 2026-09-05 · Medium confidence

Recommendation details

Best when: Reliable browser automation Claude can operate through code.

Avoid when: Avoid when the current official source does not support your required deployment, data-handling, or integration constraints.

Compatibility: Validate interfaces, data formats, and operational handoffs with the adjacent stack stages.

Prerequisites:

  • Review the current official source and implementation requirements.

Evidence: Playwright official source (2026-09-05)

Verified 2026-09-05 · Medium confidence

Implementation references

  • OpenAI Codex Workflows ↗

    Practical workflows for supplying context, reproducing defects, planning changes, and verifying results.

    developers.openai.com · documentation

Need this workflow built and running, not just linked? EE Solutions implements agentic stacks like this one. EE Solutions is a senior technology team for private capital firms and their portfolio companies.

Talk to EE Solutions ↗