One plugin for product strategy, evidence-based discovery, rapid validation, and governed AI-assisted delivery, from Accelerated Innovation.
AIPOS (AI Product Operating System) groups the work into pillars. The plugin keeps all thirteen skills together because behavior authoring, review, journey definition, and scope selection support several pillars. Each skill owns a specific action and hands off when the requested work changes.
In a terminal with Claude Code installed:
claude plugin marketplace add Accelerated-Innovation/aipos-plugins
claude plugin install aipos@aiposThe marketplace and plugin are both named aipos, in the aipos-plugins
repository (renamed from govkit-plugins; the old URL redirects). Start a new
Claude Code session after installation. For an early
installation of aipos-p1, aipos-p2, or govkit, follow the short
replacement instructions.
This repository uses Claude's .claude-plugin packaging. It does not currently
provide a Codex plugin manifest.
Describe the action and provide the relevant evidence or artifacts:
| Request | Skill |
|---|---|
| “Start a new product from this idea, or revise our strategy after these results.” | aipos-product-strategy |
| “Review the live opportunity graph for next quarter's research choices.” | aipos-exploration-planning |
| “Run a solution design canvas for this opportunity with the team.” | aipos-solution-framing |
| “Design a demand test for this opportunity.” | aipos-rapid-validation |
| “Draft acceptance criteria from these findings.” | aipos-feature-create |
| “Review these existing criteria with Product, QA, and Engineering.” | aipos-feature-refine |
| “Select the smallest useful release from these scenarios.” | aipos-feature-slice |
| “Check this reviewed repo package before implementation.” | aipos-feature-readiness |
See the complete skill catalog and prerequisites for solution framing canvases, journey authoring, dashboards, optional epics, synthetic data, and metrics.
Product strategy starts with a concise Product Opportunity Brief, even with only an idea and no graph. It connects vision, initial customers, differentiation, business viability, and outcomes to the next learning decision; research and product results return to the same brief. P1 reviews live evidence and drafts exploration decisions. P2 tests assumptions and progressively defines, reviews, and selects behavior for a commitment package. The accountable authority decides whether to approve that exact scope. P3 checks the reviewed repository package for execution readiness, then supports test data and delivery observation.
An evidence-backed opportunity can go directly to Rules and scenarios. Epics, user stories, estimates, and a decision service are optional. Existing work can enter at the appropriate stage; this is not a compulsory sequence of thirteen calls.
The workflow guide explains ownership, handoffs, and examples. Product approval, local execution readiness, current authority, and implementation evidence remain distinct. A dashboard badge or completeness score grants none of those by itself.
Read CONTRIBUTING.md. Offline tests validate packaging and the bundled scripts. Routing evaluations test selection from the descriptions; behavior evaluations test a loaded skill's coaching. Model results must be reported separately from offline test passes.
The product-strategy integration record records this addition, its verification, and the unchanged canvas handoff limits. The consolidation plan retains the earlier release history.
Licensed under MIT.