skill

superpm-skill

Single-skill PM operating system: capture intent, diverge by PM domain, draft the smallest useful artifact, verify with an independent critic. 68 pm-skills frameworks folded into one skill, supergoal-

cskwork/superpm-skillcommunitylistedunreviewed

Use skill

Set up

Skills are installed by placing their folder where your agent loads skills.

Download ZIP from GitHub Open the repository

The download comes from github.com/cskwork/superpm-skill itself, not from our server, so what you get is exactly what its author published.

Install instructions come from the README

Skills do not use an MCP config. Follow the steps in the README below.

README

synced Oct 26, 2026

/superpm

English | 한국어

One PM request in, a verified artifact out - the smallest useful artifact, checked by an independent critic against the framework and the real decision.

A single agent skill that folds the whole PM toolbox into one workflow: it captures intent from your request, diverges by PM domain into the right framework, produces the smallest useful artifact, and runs an independent critic over it before delivery - then stops.

No extra install: clone the repo, symlink it into your skills directory, then /superpm <ask>. Landing page: cskwork.github.io/superpm-skill.

What it is

phuryn/pm-skills is a marketplace of 68 PM skills across 9 plugins, each a separate file you load by name. superpm collapses that catalog into one skill, one repo. You do not pick a skill; you state the PM ask and the workflow routes it:

/superpm write a PRD for in-app referral
/superpm what's our TAM/SAM/SOM for SMB payroll in Korea
/superpm build an opportunity solution tree for activation
/superpm draft a pricing strategy for the pro tier
/superpm read this A/B test and tell me ship or kill
/superpm a GTM plan for the EU launch
/superpm review my PM resume against this job description

How it works

The same loop runs for every domain; only the framework that loads changes.

Capture - classify the ask into a PM domain + a target artifact. If genuinely ambiguous, a <=5 question interview gate fires; anything answerable from your docs/data is read, not asked.
Frame - state the real decision the artifact must serve and its completion bar.
Draft - apply the domain framework. Smallest useful artifact, real inputs, assumptions labeled.
Critic - an independent pass red-teams the draft: missing sections, unsourced claims, unvalidated assumptions, dodged decisions - each emitted as a risk.
Deliver - fold the risks back in and stop, reporting what was checked.

Domains (9 pm-skills plugins + screen-design + flow + package = 12)

The four build-planning domains form a chain on one source of truth (the 기능명세 / SPEC): EXECUTE -> FLOW -> STORYBOARD -> PACKAGE, contract in reference/model.md.

Domain	Covers
DISCOVER	ideation, risky assumptions, opportunity solution tree, feature prioritization, customer interviews, metrics
STRATEGY	strategy canvas, vision, value proposition, lean/business model, monetization, pricing, SWOT/PESTLE/Porter/Ansoff
EXECUTE	PRD, OKRs, outcome roadmap, sprint, retro, pre-mortem, stories, stakeholder map, prioritization frameworks, red-team
FLOW	user-flow graph - nodes derived from SPECs (R2), page->screen hand-off (R3), zero-dep visualization (text list + boxes)
STORYBOARD	화면설계서/기능명세 screen design docs - wireframe (plan-first) or replica/Figma, per-element spec; runs standalone via bundled HTML, optional `storyboard-spec` accelerator
PACKAGE	integrated implementation bundle - PRD+spec+flow+wireframe under one snapshotId (R6), back-link integrity, md/json/svg/html export
RESEARCH	personas, segmentation, customer journey, market sizing (TAM/SAM/SOM), competitor analysis, sentiment - grounded in live, keyless market signal (voice of customer)
ANALYTICS	NL->SQL, cohort analysis, A/B test read-out
GTM	go-to-market, beachhead, ICP, growth loops, motions, battlecard
GROWTH	marketing ideas, positioning, value-prop statements, naming, North Star metric
TOOLKIT	resume review/tailoring, NDA, privacy policy, proofreading
AI-SHIP	shipping artifacts, intended-vs-implemented gap (cited evidence)

Why a critic

PM artifacts have no unit tests, so a plausible-looking draft can be wrong. superpm borrows supergoal's discipline: an independent reader checks the artifact against three grounds - framework completeness, the real decision, and the evidence behind every claim - and names each gap as a risk before you ship it. Unsourced numbers and dodged decisions are blockers.

Install

This repo is the skill. Put it where your agent CLI finds skills:

git clone <repo-url> superpm-skill
ln -s "$(pwd)/superpm-skill" <your-agent-skills-dir>/superpm
# examples: ~/.claude/skills/superpm, ~/.codex/skills/superpm

Then in your agent CLI: /superpm <your PM ask>.

Layout

SKILL.md      thin spine: principles, intent-capture table, the loop, reference map
reference/    intent · critic · signal · model · discover · strategy · execute · flow · storyboard · package · research · analytics · gtm · growth · toolkit · ai-ship
templates/    PRD · strategy canvas · opportunity solution tree · battlecard · storyboard page/board · workspace (3-tab)
docs/         DESIGN.md

Credit

PM framework catalog: phuryn/pm-skills (MIT), curated by Paweł Huryn - merged here into one skill.
Workflow + verification discipline adapted from cskwork/supergoal-skill (MIT).
Screen-design storyboard method (optional accelerator) from cskwork/storyboard-spec (MIT).
Live-signal method (optional last30days accelerator) from mvanhorn/last30days-skill (MIT).
Keyless read techniques (Jina Reader, RSS) from Panniantong/Agent-Reach (MIT).

License

MIT. See LICENSE.