think-pdca-a3

think-pdca-a3 is a command for Claude Code from product-on-purpose/thinking-framework-skills. It costs 49 tokens per session (532 once invoked), scanned A, original, Apache-2.0.

A command that runs a structured improvement cycle: find the cause of a performance gap, choose a countermeasure, review the result, then standardize it or try again. It uses PDCA (Plan-Do-Check-Act) and the one-page A3 problem-solving format.

In plain words
What is it for?
Use it to investigate why results fall short, select an action to test, compare actual results with expectations, and decide whether to keep or revise the approach.
Why use it?
It keeps an improvement effort focused and passes only the key result from one step to the next. This avoids repeating long notes and helps turn observations into a tested action.

Command for Claude Code

Written for Claude Code: argument-hint in frontmatter.

Part of the thinking-framework-skills plugin — 68 skills, 10 commands, 1 agent shipped together

Good fit Use it to investigate why results fall short, select an action to test, compare actual results with expectations, and decide whether to keep or revise the approach.

Compare 6 commands from other repositories ↓
Install with agentmods
npx agentmods add commands/product-on-purpose/thinking-framework-skills/think-pdca-a3
Install

Getting it into your agent

One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.

Clone the repo
git clone --depth 1 https://github.com/product-on-purpose/thinking-framework-skills

Made for: Claude Code.

Or install thinking-framework-skills, the plugin that ships this one along with the rest of its 68 skills, 10 commands, 1 agent.

Wrote this? Show the measurements

A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.

agentmods badge for think-pdca-a3

README.md
[![agentmods](https://agentmods.dev/badge/commands/product-on-purpose/thinking-framework-skills/think-pdca-a3/github.svg)](https://agentmods.dev/commands/product-on-purpose/thinking-framework-skills/think-pdca-a3)
Your own site
<a href="https://agentmods.dev/commands/product-on-purpose/thinking-framework-skills/think-pdca-a3"><img src="https://agentmods.dev/badge/commands/product-on-purpose/thinking-framework-skills/think-pdca-a3/github.svg" alt="Measured on agentmods" height="20"></a>

Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.

agentmods 80×15 button for think-pdca-a3

Your own site · 80×15
<a href="https://agentmods.dev/commands/product-on-purpose/thinking-framework-skills/think-pdca-a3"><img src="https://agentmods.dev/badge/commands/product-on-purpose/thinking-framework-skills/think-pdca-a3.svg" alt="Reviewed on agentmods" width="80" height="20"></a>
Per session 49 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 532 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. A grade says what 26 rules found in the file — not that it is safe.
Origin original No closer match found in the catalogue.
Token cost

What it costs to keep this loaded

Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.

ModelPer sessionOnce invoked
Fable 5.1 $0.00049 $0.00532
Opus 5 $0.00024 $0.00266
Sonnet 5 $0.00010 $0.00106
Haiku 4.5 $0.00005 $0.00053

Measured 10d ago against content hash e9a4d6782167, method: parsed. Prices are Anthropic first-party input rates as of 2026-09-09, from the pricing page.

Security

Grade A, and why

think-pdca-a3 scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 10d ago.

A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.

Nothing flagged

None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.

commands/think-pdca-a3.md · 25 lines

What it actually says

Run the think-pdca-a3 recipe: 3 skills chained in a fixed order, carrying forward only the compressed artifact at each handoff.

The compression is the point. Do not paste a whole artifact into the next step; carry exactly what each handoff names, so the chain stays cheaper and sharper than running the skills back to back.

PDCA (Plan-Do-Check-Act, the Shewhart/Deming improvement cycle; Deming preferred Plan-Do-Study-Act) and Toyota's A3 are not a separate skill here: the reflective heart of the loop already ships as after-action-review, the forward half is execution plus a "repeat" instruction, and the A3 one-page layout is a document convention. This recipe chains the shipped moves rather than duplicating after-action-review under a more famous industrial name.

Run the steps in order, carrying forward only the compressed artifact between them:

  1. think-issue-tree (Plan, root cause) -> carry the root cause of the performance gap. Swap in think-iceberg-model when the gap is systemic (events down to structures and mental models) rather than a decomposable deviation.
  2. think-decision-option-review (Plan, countermeasure) -> carry the chosen countermeasure to test.
  3. (Do - run the change in the world; this emits no thinking artifact and the library does not own it.)
  4. think-after-action-review (Check) -> carry expected versus actual, the why, and what to sustain or change.
  5. (Act - standardize if it worked, adjust and re-run if not; a control-flow wrapper, then loop back to step 1.)

Composite artifact: an improvement-cycle record - the root cause, the countermeasure tested, the actual-versus-expected review, and the standardize-or-iterate decision. Full prose and rationale: recipes/pdca-a3.md.

Apply the chain to: $ARGUMENTS

Changes

What this file has done since we first saw it

Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.

  1. 10d ago First seen · 25 lines · 49 tokens per session scan A e9a4d6782167

Subscribe to this mod's changes

think-pdca-a3 is a command published in the GitHub repository product-on-purpose/thinking-framework-skills (15 stars, last pushed 24d ago), licensed Apache-2.0. It adds 49 tokens to every session and 532 once invoked, about $0.0002 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.