Feature spec
Goal: Send a clear invoice after finishing a job.
- Add a client, line items, a due date, and payment instructions.
- Preview the total before exporting a PDF.
- Keep an editable draft if export fails.
App Overview
AgentShell
AgentShell is an SSH client for developers who run AI coding tools like Claude Code and Gemini CLI on remote servers.
Free · In-app purchases
PRICE
Free
In-app purchases
BUSINESS MODEL
Freemium
No overall Grossing estimate available
RATING
★ 5.0
1 ratings · Top 5%
Evidence scope: No current overall US chart position is available. Category-chart positions shown in discovery are a separate scope. Revenue is a directional model estimate, not verified earnings.
Build your own version
Build Kit
Make this idea yours
A product spec, screen plans, and build steps for your coding agent.
First kit free with a verified email
Add to a Remix
Combine features and design from your favorite apps.
Kit ready. Verify your email to download it.
Generating… usually about a minute.
Make it yours Optional
What should your app do differently? Your note travels with the kit so your coding agent builds around your idea.
What's included15 files
A product spec, phased build plan, differentiation notes, and screen references.
- START_HERE
- AGENTS
- EVIDENCE
- SPEC
- BUILD_PLAN
- DIFFERENTIATION
- skills/
- screens/
Preview a sample Build Kit No email needed
Illustrative example: an invoice app for independent contractors. Your kit is tailored to the app you choose; this is not its generated content.
Screen plan
Invoice editor: Client at the top, editable line items in the middle, total and preview action at the bottom.
States: Empty draft, validation errors beside each field, exporting, and a retry action that preserves the draft.
Differentiate and validate
Hypothesis: Contractors need faster repeat invoices more than more templates.
First experiment: Test duplicating a previous job with five contractors. Watch where they hesitate before expanding the feature set.
The full kit adds a phased build plan, evidence notes, tool prompts, and implementation guidance. It is a plan for your coding agent, not a finished app.
Usage: Mechanics are fair game; never reuse the original name, branding, assets, or verbatim copy.
JavaScript is required to complete mailbox verification.
Free quick prompt For a smaller starting point
Quick prompt
Good fit for mobile-first iOS and Android prototypes with managed app scaffolding.
// RORK SYSTEM PROMPT
Rork plan preview for AgentShell: map the core Developer Tools workflow, choose the smallest differentiated feature set, define the data and monetization boundaries, then prototype the riskiest user journey first.
Preview ready. Verify your email to reveal and copy the full prompt.
JavaScript is required to complete mailbox verification. Return to this prompt preview.
Continue in Rork >
Affiliate link: CloneChart may earn a commission at no extra cost
Opportunities
AI-GENERATED OPPORTUNITY HYPOTHESES
These are generated suggestions, not verified review quotations. Check current App Store reviews and speak with users before treating a possible gap as a requirement.
Hypothesis 1
Competitive pricing with generous free tier
Assumption to check: Expensive subscription
Validate with people who use this workflow before building.
Hypothesis 2
Focus on most-requested features first
Assumption to check: Missing features
Validate with people who use this workflow before building.
Hypothesis 3
Optimize for 60fps, fast load times
Assumption to check: Poor performance
Validate with people who use this workflow before building.
App evidence
Original App Store material. Ratings and screenshots describe the existing app, not proof of demand for your version.
App Store description
AgentShell is an SSH client for developers who run AI coding tools like Claude Code and Gemini CLI on remote servers. The problem it solves You're running Claude on a dev server. It needs a yes/no answer. Your phone is in your pocket. Without notifications, you'd have no idea it's waiting. AgentShell watches your terminal for input prompts and notifies you. Works whether the app is open, backgrounded, or your phone is locked. AI agent support • Auto-launch Claude, Gemini, Cursor, Codex, or your own tools on connect • Detects when agents need confirmation or input • Push notifications show the actual question being asked • Banners when you're in another session • Per-connection working directories Lock Screen and Dynamic Island • Session status visible without unlocking • See "Working", "Waiting for input", or "Ready" at a glance • Tap to jump straight into the session Multiple sessions • Up to 10 sessions at once • Grid view with terminal previews • Sessions stay alive when you switch apps • Auto-reconnect on dropped connections • GNU Screen support Terminal • 256 colors, multiple themes • Extra keyboard row: Ctrl, Esc, Tab, arrows • Voice input • Copy/paste Security • Keys in iOS Keychain with hardware encryption • Face ID / Touch ID to access keys • Ed25519, RSA, ECDSA generation • Keys stay on device • Secure password storage Connections • Unlimited saved profiles • Quick connect to recent servers • AWS EC2 metadata support • Custom startup commands • Background connection keeper No web views Native iOS. SwiftUI. Pure Swift SSH. Launches fast, stays responsive. --- Subscription Information AgentShell Pro (Yearly) is available as an auto-renewable subscription at $9.99/year. Payment is charged to your Apple ID account at confirmation of purchase. The subscription automatically renews unless cancelled at least 24 hours before the end of the current period. You can manage and cancel your subscription in your App Store account settings. AgentShell Pro (Lifetime) is available as a one-time purchase at $24.99. Privacy Policy: https://agentshell.dev/privacy Terms of Use (EULA): https://www.apple.com/legal/internet-services/itunes/dev/stdeula/ ---
Written user reviews are not shown here. Check current App Store reviews >
Screenshots
< Scroll to see more >
Build scope
Plan a focused first version with your coding agent. These are planning assumptions, not a delivery guarantee.
Define the first version
Choose one audience and one core workflow. Use the kit to agree on its screens, data, and acceptance criteria before building.
Set boundaries
Decide which secondary features, integrations, and platform support can wait. Your version does not need to reproduce everything in the original.
Check the hard parts
Validate external services, specialist technology, data access, and ongoing costs for your chosen scope.
SUGGESTED BUILD STACK
React Native
Expo
Supabase
RevenueCat
BUILD KIT DIFFICULTY
Advanced
From the Build Kit assessment.
A reliable timeline needs an agreed scope and a technical check. Ask your agent to estimate the phases in BUILD_PLAN.md after that review.
