What does this skill do, and when should you use it?
PlanMySaaS is a free, MIT-licensed Claude Code skill that converts a single-sentence product idea into a complete SaaS blueprint. It runs an 8-stage pipeline — idea refinement, market research, product analysis, system architecture, feature specs, frontend blueprint, release plan, and a build playbook — writing one markdown file per stage into ./planmysaas-blueprint/, then an index README. The skill is plain markdown prompts with no telemetry, no signup, and no API calls. It complements the planmysaas.com dashboard, which offers 30+ pages, charts, version history, and exports at the cost of an account and credits.
Triggers on /planmysaas "idea" or natural phrasing like "I want to build a SaaS for X". It first confirms the one-line idea (asking up to two clarifying questions about audience and business model if only a domain is given), then reads each of the eight prompt files in pipeline/ sequentially and generates: 01-idea.md (refined idea, audience, business model), 02-research.md (5–8 competitors with opportunity scores, problem clusters), 03-analysis.md (business model canvas, SWOT, PMF score, risk matrix), 04-architecture.md (containers, services, data models, API surface), 05-features.md (12–20 specs with IDs, priorities, user flows, acceptance criteria), 06-frontend.md (routes, ASCII wireframes, component tree), 07-phases.md (release plan), and 08-build-playbook.md (dependency-ordered build steps with rubrics). All files are written via the Write tool into ./planmysaas-blueprint/; chat shows only 3–5 line summaries per stage; a final README.md is generated from a template plus a fixed-format summary message.
- A solo developer in Claude Code who has a fresh SaaS idea and wants research, architecture, and a feature list before writing code
- A hackathon participant needing to turn an idea into an executable build plan within a day or two
- A non-technical founder working with AI coding tools who wants a prioritized, rubric-checked build playbook to follow step by step
- Someone with a raw idea but no target audience or business model, using the skill's two clarifying questions to fill the gaps first
- Developers who prefer a fully local workflow and refuse to sign up for or upload ideas to cloud planning tools
How do you install this skill?
- This is a static source review; the skill was not executed. The pipeline/ prompt files and templates referenced by SKILL.md are absent from the evidence, so output quality and file completeness are unverified.
- Output is English-only with no Chinese support, and no mainland-China reachability notes are provided; research-stage content (competitors, market figures) depends on the Claude session and may contain directional estimates requiring manual verification.
- Publisher identity is unverified and there is no changelog; review the generated build playbook's security defaults before handing it to an AI coding tool.
- Local filesystem
Clone the repo into the Claude Code skills folder: git clone https://github.com/creationskiro/planmysaas-claude-skill ~/.claude/skills/planmysaas, then restart Claude Code or open a new session — the skill auto-discovers from ~/.claude/skills/. No build step, no dependencies.
git clone --depth 1 https://github.com/creationskiro/planmysaas-claude-skill.git ~/.claude/skills/planmysaas-claude-skillGenerated from the source repository and skill path; it copies only this skill's folder. If the author's install steps above differ, follow those first. To scope it to one project, replace ~/.claude/skills with that project's .claude/skills.
How do you use this skill?
Once installed, send your agent any of these to trigger it:
- your one-line saas idea
- AI tutor for JEE students, voice-first, ₹99/month
- turn this idea into a structured blueprint: …
In Claude Code, run /planmysaas "your one-line saas idea", e.g. /planmysaas "AI tutor for JEE students, voice-first, ₹99/month". Natural descriptions like "turn this idea into a structured blueprint: …" also trigger it. The skill creates ./planmysaas-blueprint/ in your current project and writes nine files; open README.md first, then start at Build Step 01 of 08-build-playbook.md and pass each step's acceptance rubric before advancing. If the folder already exists from a prior run, the skill asks whether to overwrite or create a -v2 folder.
What are this skill's strengths and limitations?
- Completely free under MIT — no signup, no API key, no telemetry, runs locally in your session
- Portable markdown output with zero lock-in; every stage prompt is an editable file
- The build playbook is dependency-tree ordered with 8–14 testable acceptance checks per step, well suited to driving AI coding tools incrementally
- Explicit anti-hallucination rules: invented stats must be labeled directional estimates; placeholders and TODOs forbidden
- Depth is capped by a single Claude session: the repo's own comparison says charts, 26+ artifacts, version history, and collaboration are dashboard-only features
- Research quality depends entirely on Claude's own knowledge/search; the repo ships no test suite or output-quality guarantees
- Designed and tested only for Claude Code; other Agent Skills clients require unverified adaptation
- The README's rendered screenshots come from the companion dashboard, not from skill output, so there is no directly shown example of what the markdown files actually look like
How does this skill compare with similar options?
Side by side with related skills; every score comes from the same FSRS standard.
| Skill | FS score | Stars | Last updated | License |
|---|---|---|---|---|
| PlanMySaaS this page | 55 · Use with care | ★ 12 | 6mo ago | MIT |
| i-have-adhd: ADHD-Friendly Output Mode | 65 · Recommended | ★ 57k | 14d ago | MIT |
| Product Manager Skills | 59 · Recommended | ★ 186 | 6mo ago | NOASSERTION |
| Family History Research Planning Skill | 56 · Use with care | ★ 117 | 10mo ago | MIT |
| Pricing Advisor (The Minimalist Entrepreneur Pricing Skill) | 55 · Use with care | ★ 11k | 6mo ago | — |
The README explicitly compares the skill against its sibling product, the planmysaas.com dashboard: the skill is free, account-free, editor-native, and outputs 9 markdown files; the dashboard requires an account and network, uses credits (100 free), and adds 30+ page blueprints, inline charts, a visual architecture canvas, version diffing, per-stage re-runs, team workspaces, a kanban build tracker, and PDF/JSON/ZIP exports. Use the skill for exploration and fast loops; use the dashboard for real, long-running team projects.
How did FollowSkills review this skill?
Pure markdown, no network calls, no dependencies; explicitly read-only on user files and scoped to ./planmysaas-blueprint/, with an overwrite/v2 confirmation rule. Clear scope and rollback. Deducted: referenced pipeline prompt files are absent from the evidence so the injection surface cannot be statically excluded, and there is no active check on generated content safety. Full marks not awarded.
Orchestration is self-consistent, paths fixed, per-stage fresh prompt reads; happy path plausible on paper. Deducted: the 8 pipeline prompt files and templates the skill depends on are not present in the reviewed evidence, so key paths cannot be verified statically; no tests and thin handling of abnormal inputs. Static cap 10.
Clear trigger phrases, dual slash/natural-language entry, well-defined output contract, honest 'no fabricated numbers' rule. Deducted: no declared non-fit boundaries, English-only saved output with explicit non-English restrictions, no Chinese support, and no statement on mainland-China reachability.
Well-layered docs: README with install/customise/roadmap, CONTRIBUTING with explicit PR standards, SECURITY.md with disclosure process, clear MIT license. Deducted: no CHANGELOG, versioning only a v1.0 mention, publisher unverified, maintenance rests on a single author.
Claims 9 structured files in ~2 minutes; rules forbid placeholders and fabricated stats; README screenshots indirectly show the target output. Deducted: static review cannot verify generation quality — value depends entirely on Claude following prompts; screenshots come from the paid dashboard, not the skill's own output, and comparative benefit is asserted marketing-wise. Static cap 7.
Claims of no telemetry and no network are consistent with the sources; SECURITY.md's attack-surface analysis is reasonable. Deducted: no third-party execution evidence, no CI, no committed tests or real output samples; claims like real competitor research are independently unverifiable and marketing is interleaved with fact.
Open a dimension to read why it scored that way
Evidence confidence：Low — Mostly static review, author material or a limited demo; useful for discovery, not high-risk decisions.
See the full review method →