Explore skills

Converts raster images (photos, paintings, illustrations, line art) into SVG vector reproductions. Use to vectorize, trace, or convert images to SVG, or to faithfully reproduce artwork as vector graphics.

Universal environment variable loader for AI agent environments. Loads secrets and config from Claude.ai, Claude Code, OpenAI Codex, Jules, and standard .env files.

Installs skills from github.com/oaustegard/claude-skills into /mnt/skills/user. Use when the user mentions install, load, add, update, refresh skills, or references a skill not currently installed.

Executes programs on a compiled transformer stack machine where every instruction fetch and memory read uses a parabolic attention head. It demonstrates that transformer attention + FF layers can implement a working computer.

Discover and load skills on demand from /mnt/skills/user/. This tool helps you find capabilities, get full skill descriptions, or browse the catalog using 'list', 'search', and 'show' commands.

Dispatched after a task lands at `in_review` to produce a structured CTO-grade verdict on the work and its PR. Two invocation paths: composer Phase 4 (orchestrator dispatches after the implementer's `in_review` write, surfaces the verdict to HOTL, stops), and direct mode from the piyaz skill on requests ("review VF-N", "review this PR", "review <PR URL>"). Reads `piyaz_context depth='review'` for

Dispatched after a task lands at `in_review` to produce a structured CTO-grade verdict on the work and its PR. Two invocation paths: composer Phase 4 (orchestrator dispatches after the implementer's `in_review` write, surfaces the verdict to HOTL, stops), and direct mode from the piyaz skill on requests ("review VF-N", "review this PR", "review <PR URL>"). Reads `piyaz_context depth='review'` for

Generates, validates, and outputs new ideas based on existing knowledge, combining combinatorial creativity, cross-domain analogy, and minimum experiments. Use when the user wants fresh ideas, new product concepts, or creative solutions.

Load an ALIVE walnut -- read kernel files, show current state, surface one observation, ask what to work on

Morning briefing -- read all walnut states, surface priorities, inbox count, stale walnuts, people nudges

The human mentions a walnut to work on, asks about a specific venture/experiment/project, or wants to check status — not just explicit 'load X'. Load the brief pack (3 files), resolve the people involved, check the active bundle — then surface one observation and ask what to work on. Context loads in tiers: walnut and people are automatic, bundle depth is offered.

Morning operating system -- sync inputs, read everything, surface priorities, show the day