# Run a plan-then-build method with subagents in a coding harness

> Run a plan-then-build method with subagents in a coding harness: 0 companies fully do it, 1 with a limit, 78 carry the tag. Every listing in Code & developer platforms with the reason and a source.

- Canonical: https://www.anchorterminal.com/niches/code.dev-method/
- Markdown: https://www.anchorterminal.com/niches/code.dev-method/index.md (~1,250 tokens)
- Slim: https://www.anchorterminal.com/niches/code.dev-method/index.min.md (~280 tokens, same facts, less prose, for token-sensitive contexts)
- JSON: https://www.anchorterminal.com/niches/code.dev-method/index.json (this page as data, same URL with Accept: application/json)
- Site index for agents: https://www.anchorterminal.com/llms.txt (the key docs and every listing in one file, at most 16 MB, about 4 million tokens: https://www.anchorterminal.com/llms-full.txt)
- API: https://www.anchorterminal.com/api/v1/index.json
- MCP: https://www.anchorterminal.com/mcp (the whole catalogue as tools, no key)
- DNS: dig +short TXT <slug>.grade.anchorterminal.com (the grade as a DNS TXT record, for a sandbox with no HTTP where names still resolve; also top., best.<category>. and <a>.vs.<b>. under the zone, and the same under grade.letme.dev)
- Licence: CC BY 4.0, credit Anchor Terminal and link this page (https://www.anchorterminal.com/terms/#data)
- Updated: 2026-11-15

No company's docs show a full fit yet. The listings below do the job with a limit, or not at all.

| Companies, full fit | With a limit | Carry the tag | Share of the category's work | First-hand accounts |
| --- | --- | --- | --- | --- |
| 0 | 1 | 78 | 8 | 6 |

## What counts as done

A design document and plan exist before code, each plan task ends with a test and a commit, the harness invoked the skill at the right step (visible in the transcript), and the final summary states what was verified and what was not.

Measured, for our own tests: each plan task's test passes.

How agents ask for it: brainstorm this feature then write a plan; use subagent-driven development; finish this branch and open the PR; review what the subagent did before we merge

## Does the job, with a limit

It does the job, but something in the docs or a user's account will get in an agent's way. The limit is named in each row.

- **Superpowers (Prime Radiant (no legal form or registration found on its pages); the MIT LICENSE names Jesse Vincent)**: brainstorming, writing-plans, subagent-driven-development, test-driven-development and verification-before-completion exist with installs for 16 harnesses, and a diagnosing-superpowers skill; but two open issues show no skill firing and completion claimed before verification, and each harness needs its own install.

- not this job: amCharts 5 MCP Server (amCharts): Not a method or skills library.
- not this job: Azure DevOps MCP Server (Microsoft Corporation): Not a method or skills library.
- not this job: Context7 (Upstash): Not a method or skills library.
- not this job: Entire (Entire Inc): Not a method or skills library; sessions are captured, not steered.
- not this job: Git (MCP reference server) (Model Context Protocol, a Series of LF Projects, LLC): Not a method or skills library.
- not this job: GitHub MCP Server (GitHub, Inc.): Not a method or skills library.
- not this job: GitLab MCP Server (GitLab Inc.): Not a method or skills library.
- not this job: GitMCP (Ido Salomon (GitMCP open-source project)): Not a method or skills library.
- not this job: graft (Vishisht Choudhary (individual)): Not a method or skills library.
- not this job: Graphify (Graphify Labs): Not a method or skills library.
- not this job: Jam MCP (Jam): Not a method or skills library.
- not this job: Microsoft Learn MCP Server (Microsoft Corporation): Ships skills that tell agents when to call its tools, not a build method.
- not this job: ModusToolbox AI Assistant (Infineon Technologies AG): Not a method or skills library.
- not this job: Momentic MCP server (Momentic): Not a method or skills library.
- not this job: REA (morluto): Not a method or skills library.
- not this job: Ref (Ref Software, Inc): Not a method or skills library.
- not this job: Salesforce DX MCP Server (Salesforce, Inc.): Not a method or skills library.
- not this job: Sourcegraph MCP Server (Sourcegraph, Inc.): Not a method or skills library.
- not this job: TokenEconomizer (Microsoft Corporation (repository in the microsoft GitHub organisation with Microsoft's SECURITY.md); co-authors at UC Berkeley): Prunes tools, injects facts and redirects stuck turns; it has no plan, test or commit steps, and supports ten-patch version ranges per harness.
- not this job: Void (Cloudflare, Inc. (VoidZero)): Skills point agents at void commands; not a build method.

## What stops agents finishing

- skills listed as available but never invoked, and completion claimed before verification
- bootstrap injected on every session start tells the model to invoke skills before simple replies
- a second harness needs a second install
- supported agent version ranges of ten patch releases (TokenEconomizer)

## Around it

Before: agent-harnesses.pair-in-terminal. After: code.ticket-to-pr, code.review-pr.

## Where this comes from

Weight 8 of 100 in Code & developer platforms: 6 of 65 accounts, all Superpowers (Claude Code, Codex, OpenCode, Pi, metaswarm); TokenEconomizer had none

As data: https://www.anchorterminal.com/api/v1/use-cases/code.dev-method.json, or the MCP tool get_use_case with id code.dev-method. Draft, read 2026-11-12 from public docs and first-hand accounts, not yet reviewed by a person.

---

Query this catalogue directly over MCP: https://www.anchorterminal.com/mcp (Streamable HTTP, no key). How: https://www.anchorterminal.com/agents/#mcp
