erpaval
Workflow for autonomous software development with Claude Code. Classifiers route scope, complexity, directory state, and spec readiness before committing to…
Optional ask, prior memo path, or challenge statement.
$ npx -y skills add theagenticguy/erpaval --skill product-strategy --agent claude-codeHow it fires
How this skill gets triggered: by you, by Claude, or both.
/product-strategyContext preview
The summary Claude sees to decide when to auto-load this skill.
Optional ask, prior memo path, or challenge statement.
name: product-strategy
description: >
File-first orchestrator for business and product strategy frameworks.
Frames the challenge with a CPO / CSO / VP Design role, spawns parallel
framework agents (Rumelt kernel + Crux, Wardley Maps, Minto Pyramid as
thinking structure, Working Backwards PR-FAQ as discovery artifact),
synthesizes outputs into one strategy memo, runs a critic pass. Ships
per-framework references, leadership role prompts, templates for
strategy memos, PR-FAQs, Wardley maps, Minto arguments. Use when the
user asks to do strategy work, run a Rumelt kernel, build a Wardley
map, decide build-vs-buy, articulate the crux of a challenge, draft a
PR-FAQ as discovery, write an executive strategy memo, structure an
argument with Minto Pyramid, think from the CPO / CSO / VP Design
perspective, or mentions Good Strategy Bad Strategy, The Crux, Rumelt,
Simon Wardley, SCQA, or strategy thesis. Do NOT use for PRDs or HMW
framing — use product-discovery.
arguments:
- name: source
description: Optional ask, prior memo path, or challenge statement.
required: false
user_facing: true
metadata:
libraries: []| Reference | When to load | | ------------------------------------------- | ----------------------------------------------------------------------------------------- | | `references/orchestrator.md` | Running the pipeline — five phases plus a critic pass | | `references/write-protocol.md` | Canonical write-protocol block copied verbatim into every subagent prompt | | `references/frameworks/INDEX.md` | Routing guide — decision table + cross-framework composition | | `references/frameworks/*.md` | Per-framework files — Rumelt kernel, Wardley maps, Minto pyramid, Working Backwards | | `references/roles/cpo.md` | CPO role — product-line framer | | `references/roles/cso.md` | CSO role — portfolio framer (build-vs-buy, M&A, 3–10 year horizon) | | `references/roles/vp-design.md` | VP Design role — customer-archetype and experience framer | | `references/roles/rumelt-architect.md` | Runs the three-part kernel plus The Crux | | `references/roles/wardley-cartographer.md` | Builds the text-form Wardley map | | `references/roles/pr-faq-discovery.md` | Drafts a PR-FAQ as a discovery artifact (forcing function for customer-value clarity) | | `references/roles/minto-pyramid-builder.md` | Composes SCQA + MECE argument pyramid | | `references/roles/strategy-synthesizer.md` | Reads every packet, composes the final strategy memo | | `references/roles/strategy-critic.md` | Reviews the memo on one axis (coherent + defensible), multi-dimensional rubric | | `templates/worklog-skeleton.md` | Per-role work-log skeleton with embedded write-protocol | | `templates/strategy-memo.md` | Final deliverable skeleton — Executive Summary, Diagnosis, Guiding Policy, Actions, Risks | | `templates/rumelt-packet.md` | Rumelt architect's output shape | | `templates/wardley-packet.md` | Wardley cartographer's output shape | | `templates/pr-faq-discovery.md` | PR-FAQ skeleton — press release + external FAQ + internal FAQ | | `templates/minto-outline.md` | SCQA + MECE outline |
One skill, one pipeline. Takes a strategic question → polished strategy memo through five phases that write their outputs to disk in real time. Framework agents in Phase 2 run in parallel. Synthesis and critic pass happen in sequence. The grouped framework file loads on demand — its decision table routes the orchestrator without every framework body being read.
flowchart LR
A[Phase 0<br/>Route] --> B[Phase 1<br/>Frame<br/>CPO / CSO / VP-Design]
B --> C[Phase 2<br/>Parallel framework agents]
C --> C1[Rumelt architect]
C --> C2[Wardley cartographer]
C --> C3[PR-FAQ discovery]
C --> C4[Minto pyramid builder]
C1 --> D[Phase 3<br/>Synthesize<br/>strategy-synthesizer]
C2 --> D
C3 --> D
C4 --> D
D --> E[Phase 3.5<br/>Critic review<br/>strategy-critic]
E --> |Revise up to 2 rounds| D
E --> F[Phase 4<br/>Deliver + handoff]Every role runs as a general-purpose `Agent` with `model: "opus"` and a prompt built from the matching file in `references/roles/`. Phase 2 fans out in parallel with `run_in_background: true`. Full runbook with prompts, check-in cadence, and stuck detection lives in `references/orchestrator.md`.
| User signal | Fan-out | | ------------------------------------------------------------------------ | --------------------------------- | | "Run a Rumelt kernel" / "what's the crux" / "our strategy feels fluffy" | rumelt-architect only | | "Build a Wardley map" / "should we build or buy X" / "platform strategy" | wardley-cartographer only | | "Draft a PR-FAQ for a new product idea" | pr-faq-discovery only | | "Structure this argument" / "build a Minto py
Autonomous software development for Claude Code. ERPAVal stands for Explore · Research · Plan · Act · Validate — the five-phase loop, plus a sixth Compound phase that writes durable lessons to disk so the next session inherits what this one learned.
Repo: theagenticguy/erpaval
Workflow for autonomous software development with Claude Code. Classifiers route scope, complexity, directory state, and spec readiness before committing to…
Shared canonical references for product-design frameworks — Pyramid Principle, Working Backwards, Double Diamond, research design, methodology selection.…
Optional path to raw content (brief, notes, tickets) or a one-line topic.
Agent UX patterns and human-in-the-loop design: Twilio A2H protocol, Levels of Autonomy (L0-L5), inbox pattern, progressive trust, decision journal, gate…
Customer research and problem framing using the research-design discipline: hypothesis, null hypothesis, MECE questions, methods, findings. Produces…
Audit and rewrite prompts against Claude Opus 4.7 instruction-following best practices. Produces a scored audit plus a rewritten prompt. Use when the user asks…