Skip to main content

Skyler HaliskyTechnical SupportAI-assisted Builder

The Work: 5 deliverables

The Work

A small body of AI-assisted tools and reference materials. Built slowly, documented honestly, sized to a single careful hand.

Deliverables

  • On the App Store · US and Canada · Sept 2026

    Verified

    Flagstone

    Community map of accessibility barriers: report a broken ramp, verify a neighbour's, see it marked fixed. Privacy-first: no ads, no analytics, no trackers.

    View project Live GitHub
    Flagstone's Explore map with an expanded verified Steep grade report, a real barrier photograph, orange map pins, map controls, and bottom navigation.
    Flagstone's Explore map with an expanded verified Steep grade report, a real barrier photograph, orange map pins, map controls, and bottom navigation.

  • In active use, on my projects only

    Verified

    Claude Corp

    A governed multi-agent development system I designed and operate: 15 AI roles, a written Constitution, review gates, and autonomy bounded by explicit rules.

    View project Live
    The Claude Corp front door, with the fifteen-role pipeline under the editorial hero.
    The Claude Corp front door, with the fifteen-role pipeline under the editorial hero.
    The Claude Corp front door on a phone, with the editorial hero and the governed agent pipeline below it.
    The Claude Corp front door on a phone, with the editorial hero and the governed agent pipeline below it.

  • Private operator app · public synthetic demo

    Verified

    Claude Corp Dashboard

    The operator console for my agent team: what needs me, what is blocked, how each project is doing. Private in use; a public demo runs on synthetic data only.

    View project Live
    The Dashboard command center, showing project health, agent activity and pending decisions on one calm board.
    The Dashboard command center, showing project health, agent activity and pending decisions on one calm board.
    The Dashboard command center on a phone, showing project health, agent activity, and pending decisions in a focused view.
    The Dashboard command center on a phone, showing project health, agent activity, and pending decisions in a focused view.

  • Live and evolving · no backend of mine

    Verified

    Prompt Library

    Prompts as reusable assets, not scattered text: organize, search, fill variables, run. Kept in your browser; each run goes straight to Anthropic with your key.

    View project Live GitHub
    The Prompt Library home, with search and category chips over the prompt grid and the featured card leading.
    The Prompt Library home, with search and category chips over the prompt grid and the featured card leading.
    The Prompt Library home on a phone, with search, category chips, and prompt cards in a readable single-column view.
    The Prompt Library home on a phone, with search, category chips, and prompt cards in a readable single-column view.

  • Complete · live · no backend

    Verified

    Ghost Code

    A terminal-command trainer and my first real build, made to get fluent in the command line and Claude Code. 56 cards, Arcade and Learn modes, no dependencies.

    View project Live GitHub
    A Ghost Code round in play, with the Phantom at the centre of four command tokens and the prompt above.
    A Ghost Code round in play, with the Phantom at the centre of four command tokens and the prompt above.
    A Ghost Code round on a phone, with the Phantom, four command tokens, prompt, and score in a focused play view.
    A Ghost Code round on a phone, with the Phantom, four command tokens, prompt, and score in a focused play view.
Back to home