create-swe-project

create-swe-project is a skill for Claude Code from fledgeling-co/fledgeling-plugins. It costs 355 tokens per session (5,593 once invoked), scanned B, original, MIT.

A project-starting guide and scaffold for turning an app or product idea into a working new project in the /Dev directory. It initializes the project, creates its native files, and applies the project's conventions.

In plain words
What is it for?
Use it when starting a new app, product, prototype, or other greenfield project. It helps make design and planning files, initialize version control, and prepare the chosen project technology.
Why use it?
It gives a new project a consistent starting structure and setup, reducing the manual work needed before development can begin.

Skill for Claude Code

Written for Claude Code: shipped in a Claude Code plugin. Also seen: reads .claude/ paths; mentions CLAUDE.md; names the AskUserQuestion tool.

Part of the create-swe-project plugin — 1 skill shipped together

Install

Getting it into your agent

One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.

agentmods
npx agentmods add skills/fledgeling-co/fledgeling-plugins/create-swe-project
Any agent
npx skills add fledgeling-co/fledgeling-plugins --skill create-swe-project
Clone the repo
git clone --depth 1 https://github.com/fledgeling-co/fledgeling-plugins

Made for: Claude Code.

Or install create-swe-project, the plugin that ships this one along with the rest of its 1 skill.

Wrote this? Show the measurements

A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.

agentmods badge for create-swe-project

README.md
[![agentmods](https://agentmods.dev/badge/skills/fledgeling-co/fledgeling-plugins/create-swe-project.svg)](https://agentmods.dev/skills/fledgeling-co/fledgeling-plugins/create-swe-project)
Your own site
<a href="https://agentmods.dev/skills/fledgeling-co/fledgeling-plugins/create-swe-project"><img src="https://agentmods.dev/badge/skills/fledgeling-co/fledgeling-plugins/create-swe-project.svg" alt="Measured on agentmods" height="20"></a>
Per session 355 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.
When invoked 5,593 The whole file, excluding the scripts and references it only reads on demand.
Security scan B 1 finding. Scan, not verified.
Origin original No closer match found in the catalogue.
Token cost

What it costs to keep this loaded

Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.

ModelPer sessionOnce invoked
Fable 5.1 $0.00355 $0.05593
Opus 5 $0.00178 $0.02797
Sonnet 5 $0.00071 $0.01119
Haiku 4.5 $0.00036 $0.00559

Measured 5d ago against content hash e9789f78d4fa, method: parsed. Prices are Anthropic first-party input rates as of 2026-09-06, from the pricing page.

Security

Grade B, and why

create-swe-project scanned grade B with 1 finding against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 5d ago.

The scan reads SKILL.md. This mod also ships 25 executable files (scripts/canary.sh, scripts/doctor.sh, scripts/drift.sh, …), listed below but not scanned — reading those needs a real analyzer, not pattern matching.

A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.

Asks for rootmediumPrivilege escalation

A mod that escalates privileges can change anything on the machine, not only the project.

- Walk the user through `SETUP-NEXT-STEPS.md` — the only manual parts: the `/etc/hosts` + Caddy `conf.d` mirror commands (need sudo — suggest they run them with the `!` prefix), `vercel link`, `DEVELOPMENT_TEAM` for rele
plugins/create-swe-project/skills/create-swe-project/SKILL.md · 116 lines

How it starts

The opening of the file, as written. The whole thing — 116 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Slipway — launch a new project

Take a project idea to a completely set-up, working project in ~/Dev/<codename>: gate-green, git-initialised, native projects generated, conventions wired. The trick is that scripts/scaffold.sh + templates/ already contain everything common between projects (distilled from zephyr, perch, fledgeling-app, dAIolog and the team-files operating specs) — so your job is the decisions, and the script's job is the files. Do not hand-write files the templates cover; that is what makes setup cheap.

Running as a Gemini model? Read gemini.md in this directory first, then follow this file with the overrides it names. It converts §5's launch-pipeline phases into a chain of files (DESIGN.md, UX.md, INDEX.md) with a filled mock-inventory ledger, a bound ledger read back off the built marketing page, and a route-out for the static-page and brownfield work the corpus measures this family collapsing on. Other models skip it.

0 — Doctor

Run <skill-dir>/scripts/doctor.sh (read-only JSON) before the interview so recommendations match the machine: no macos/ios module without xcodegen, no 1Password seeding without op, no rust without cargo. Mention anything missing that the chosen modules would need.

1 — Understand the idea, generate codenames

From the user's idea, decide which modules fit and generate three codename options: single word, lowercase, evocative rather than descriptive — the portfolio precedent is zephyr, perch, anvil, loupe, tare, sift, margin. Check ls ~/Dev first and never propose a name that collides with an existing directory.

2 — Ask the clarifying questions up front (one round)

One AskUserQuestion call, all questions together, before anything is created:

  • Codename — the 3 generated options (put your recommended one first, marked "(Recommended)"); the built-in Other lets them type their own. Show what each implies: ~/Dev/<codename>, <codename>.local.
  • Modules (multiSelect) — web Next.js app · api NestJS API · macos SwiftUI app · ios SwiftUI iOS app · rn React Native (Expo) mobile app · tokens design-tokens package · data Mongoose+Redis layer · auth sign-in + emails + dev login · admin allowlisted admin console · push APNs notifications · waitlist referral-loop pre-launch waitlist · observability Sentry wiring (inert until DSN) · rust cross-OS core crate. Infer before asking: a product with accounts implies auth (which implies data); "admin", "back office", "moderation" implies admin; a native/mobile app with alerts implies push; a SaaS gets web+tokens+data+auth by default; a marketing site gets web+tokens only; an unvalidated/pre-launch idea adds waitlist. Put the inferred set in the options as the recommended choice and let the user adjust — ask rather than guess only when the idea genuinely doesn't settle it. Implications are automatic in the script (push→auth→data→web; admin→data).
  • Mobile approach (only when the idea implies a mobile app and the modules answer hasn't settled it) — native Swift/SwiftUI (ios, XcodeGen, simulator-first) vs React Native/Expo (rn, one codebase for iOS+Android, hoisted pnpm + monorepo Metro per BP §18). Recommend Swift when the app is Apple-only or needs deep platform integration; Expo when Android matters or the team iterates fastest in React.
  • macOS style (only when macos selected) — window (Recommended): native NavigationSplitView shell with sidebar + top nav with search · menubar: MenuBarExtra agent app, no Dock icon (LSUIElement). Passed as --macos-style window|menubar.
  • Bundle-id prefix (only when macos/ios/rn selected) — app.<codename> (Recommended) · dev.<codename> · com.<codename> · custom. Apps get .mac / .ios / .mobile suffixes — distinct identity per sibling app from day one (BP §18).
  • GitHub org — where the repo will live: fledgeling-co (personal-product studio) · lprhodes (experiments/personal) · Diolog26 / DiologIR (Diolog work) · none for now. Infer from the idea (a Diolog feature → Diolog org; a product → fledgeling-co; a quick experiment → lprhodes) and put the inferred org first. Passed as --github-org: the remote is set and the gh repo create --push command lands in next-steps — nothing is pushed without the user running it.
  • macOS distribution (only when macos selected) — direct (Recommended): Developer ID + notarized DMG, no sandbox · mas: Mac App Store / TestFlight, App Sandbox on (a review requirement). Passed as --macos-dist; it selects the entitlements file, so it must be asked up front, not retrofitted.
  • Design reference (only when a UI surface is selected) — a site or brand to bootstrap DESIGN.md from (via design-md-from-website, which measures real computed styles), or skip to design fresh later. Passed as --design-ref; recorded in the manifest and next-steps.
  • 1Password — which account/team id and vault this project's secrets live in. Run op account list and op vault list --account <id> first and present the real values as options (plus "skip"). The chosen pair is passed as --op-account / --op-vault and seeded into apps/web/.env.local (identifiers only — actual secret values are resolved later by scripts/env-pull.sh via the op CLI, straight into the file, never through your context).

Read the full file on GitHub · 116 lines

Files

What ships with it

60 files beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.

Changes

What this file has done since we first saw it

Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.

  1. 5d ago First seen · 116 lines · 355 tokens per session scan B e9789f78d4fa

Subscribe to this mod's changes

create-swe-project is a skill published in the GitHub repository fledgeling-co/fledgeling-plugins (2 stars, last pushed today), licensed MIT. It adds 355 tokens to every session and 5,593 once invoked, about $0.0018 per session on Opus 5. A static security scan graded it B with 1 finding (asks for root). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.

Related

Other skills, from other repositories

systematic-debugging

Use when encountering any bug, test failure, or unexpected behavior, before proposing fixes.

obra/superpowers · 21 tokens

local-ai-agents

Build local-first AI agents that run entirely on a developer workstation with Microsoft Foundry Local and Qwen function-calling models. Covers Small Language Models (SLMs), the OpenAI-compatible local endpoint, sandboxed local tools, local RAG with Chroma, local MCP servers, hybrid cloud/local routing, and the…

microsoft/ai-agents-for-beginners · 200 tokens

next-cache-components-adoption

Turn on Cache Components in a Next.js app and resolve the blocking routes it surfaces. Use when the user wants to enable, adopt, or migrate to Cache Components, flip the cacheComponents flag, work through a flood of blocking-prerender / instant validation errors, run the cache-components-instant-false codemod, or…

vercel/next.js · 95 tokens

next-partial-prefetching-adoption

Turn on Partial Prefetching in a Next.js app and work through the insights it surfaces. Use when the user wants to enable or adopt Partial Prefetching, flip the partialPrefetching flag, opt routes in with export const prefetch = 'partial', audit Link prefetch={true} behavior, preserve existing prefetched UI with…

vercel/next.js · 103 tokens

chronicle

Analyze Copilot session history for standup reports, usage tips, session search, and session reindexing. Use when the user asks for a standup, daily summary, usage tips, workflow recommendations, wants to search or find past sessions by keyword/file/PR, wants to reindex their session store, or asks about deleting…

microsoft/vscode · 72 tokens

imagegen

Generate or edit raster images when the task benefits from AI-created bitmap visuals such as photos, illustrations, textures, sprites, mockups, or transparent-background cutouts. Use when Codex should create a brand-new image, transform an existing image, or derive visual variants from references, and the output…

openai/codex · 113 tokens