mobilekit
Skill1 starsChanged 45 days agoMITView on GitHub
Guided end-to-end workflow for building and modernizing React Native / Expo mobile apps — product discovery, design conception, implementation, store release, and post-release observability. Use when starting a mobile app, deciding what to build next in an Expo project, modernizing a legacy or hybrid mobile app, adding one capability to an existing app (auth, payments, push, offline, AI, store release…), or on any /mobilekit command.
In this repository11 other files it gives its agents.
- mobilekit/AGENTS.mdAGENTS.md
- alteregoSkill
- cluster-analyzerSkill
- devops-expertSkill
- doc-digestSkill
- dossie-politicoSkill
- forjaSkill
- guardrailsSkill
- lapidaSkill
- mentatSkill
- performance-architectSkill
The file
We split this skill into the rules it gives an agent
3 rules. About 0 are common, 0 are distinctive, and 3 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈0
Distinctive · estimate
This author's own choices. Worth a look.
3
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Very specific: line 59
Read prompts/RULES.md, docs/PRODUCT.md, docs/DESIGN.md (if present) and the project's AGENTS.md.
- Very specific: line 60
Read the prompt and follow its ## Prompt section as written — it is the instruction, not a summary to paraphrase.
- Very specific: line 64
If docs/BUILD-PLAN.md exists, mark that step done in it.
≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
mobilekit
A phase-driven workflow over a prompt library. Everything needed to run it lives in this skill: workflow/ holds one file per phase, prompts/ holds the library those phases execute.
The value over reading the files directly is that it knows where the project is and refuses steps whose prerequisites are missing.
Read first, always
prompts/RULES.md— how prompts are run: grill one question at a time, look facts up, request documentation, invent nothing, report honestly. It is the single source of truth; nothing here restates it.docs/PRODUCT.mdin the target project — it gates everything. Absent → the answer to almost any request is "run discovery first".
The phases
Each is a file in workflow/. Read the one you are running; skip the rest.
| Phase | File | Produces |
| Install into a project | workflow/init.md | AGENTS.md, docs/ |
| Product discovery | workflow/discovery.md | docs/PRODUCT.md |
| Design conception | workflow/design.md | docs/DESIGN.md |
| Cut the plan | workflow/plan.md | docs/BUILD-PLAN.md |
| Run the next step | workflow/next.md | ticks a box |
| Run one step out of order | workflow/build.md | — |
| Release gate | workflow/ship.md | a submitted build |
| Where the project is | workflow/status.md | nothing |
| Watch production | prompts/8-observability/post-release.md | the four numbers |
| Modernize an existing app | prompts/9-maintain/legacy-modernization.md | docs/MODERNIZATION.md |
| Market contact per phase (optional) | prompts/1-discovery/market-signal.md | docs/LAUNCHES.md |
| Decide the next cycle after release | prompts/8-observability/next-feature.md | appends to docs/BUILD-PLAN.md |
Slash commands, where the host has them, are thin aliases: /mobilekit:next → workflow/next.md. Where it has none, read the phase file directly — that is the whole mechanism.
A legacy app enters through modernization, not discovery. An app that already ships has a domain; the job is to inventory it and choose a migration path, then join the normal flow.
State
The presence of three documents is the phase. Never guess at a missing one.
| File | Written by | Means |
docs/PRODUCT.md | discovery | The product is defined. Nothing else runs before this exists. |
docs/DESIGN.md | design | Screens, navigation and per-screen states are decided. |
docs/BUILD-PLAN.md | plan | The ordered checklist of only the steps this app needs. The progress tracker. |
Executing a prompt
Prompts share a shape: a header saying when it applies, a ## Prompt section, a ### Grill block holding the developer's decisions, and a ### Done when checklist. The ones that produce a document rather than code — product-discovery.md, monetization.md and design-conception.md — carry an ## Output structure in place of the checklist, and are a grill from end to end.
- Read
prompts/RULES.md,docs/PRODUCT.md,docs/DESIGN.md(if present) and the project'sAGENTS.md. - Read the prompt and follow its
## Promptsection as written — it is the instruction, not a summary to paraphrase. - Work its
### Grillblock one question at a time, recommending an answer to each. Look facts up; put decisions to the developer. - Build.
- Walk
Done whenand report each item met, not met, or not verifiable. - If
docs/BUILD-PLAN.mdexists, mark that step done in it.
Finding a prompt
prompts/ is organized by phase: the numbered folders are the phase order, filenames are the ids. The index — every prompt with its description — and the reference build order live in prompts/README.md.
Where the library is read from
Prompts are read from this skill, wherever it is installed. workflow/init.md writes documents into the project and copies nothing — a correction to a prompt reaches every project at once. A project needing its own rules writes docs/mobilekit-overrides.md, read after prompts/RULES.md and winning on conflict.
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
Reviews can't be read right now.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.