Development

Give a site its icons as families, not one drawing at a time - inventory every icon on the built pages beside the copy it sits with (`pnpm kit icons audit`), pick a family per kind with the owner (line icons from Lucide, other companies' marks from Simple Icons, the site's own marks generated from primitives), add them with `pnpm kit icons add` or render them with `pnpm kit icons family`, and prove the change. Use when an icon "looks like a template", "means nothing", is inconsistent with its...

Draw the site's own graphics as SVG in the lab - an icon beyond the families, a mark beside copy, an illustration, an animated logo or brand mark, a short 2D loop - starting from the artwork the site already has, looked at on the site's tokens in both schemes, at the sizes it ships and enlarged, and on a phone (`pnpm kit lab serve`), checked frame by frame and as the embedded image, shipped pre-rendered as the file a page embeds or as inline Icon data (`pnpm kit lab render`, `pnpm kit icons a...

Design a site built on agentic-cms, from the wireframe to a look of its own or from one look to the next - tokens first, then the chrome, then the sections - one element at a time through candidates the owner picks from, a build on the dev server the owner looks at, and a pixel proof before anything merges. Use when asked to design, restyle, theme, brand, redesign or "make it look like" anything on the site, or to start designing a site from scratch.

Prove a change on the site's pixels with the screenshot harness - a baseline of the commit it starts from, a capture after, a compare read as a report - so a refactor shows it moved nothing and a design change shows it moved only what it meant to. Use before any merge that touches src/, styles or images, when asked to prove, verify, check parity or "make sure nothing else changed", and after a `SIZE`, `CHANGED` or `reflow` line needs reading.

Run a design round the way it works - the owner says what they want or what looks wrong, the agent brainstorms three to five ideas in one message, the owner picks which to build, the agent builds them as real components on a throwaway demo route (`pnpm kit demo new`) in the section's real frame with the page's real copy on the site's own theme, the current version last, the owner looks on the dev server and picks by letter or edits the pick in words, and on the owner's word the winner becomes...

Turn a screenshot claim ("unaligned", "too big", "appears too early", "looks broken", "cut off") into numbers before editing anything - the element's box, computed styles, stacking contexts, reveal state and timeline from `pnpm kit probe`, a crop from `pnpm kit shot` - so the fix follows the measurement. Use whenever the owner sends a screenshot or describes what a page looks like, and before touching any layout, spacing, size, z-index or timing.

Give a site its icons as families, not one drawing at a time - inventory every icon on the built pages beside the copy it sits with (`pnpm kit icons audit`), pick a family per kind with the owner (line icons from Lucide, other companies' marks from Simple Icons, the site's own marks generated from primitives), add them with `pnpm kit icons add` or render them with `pnpm kit icons family`, and prove the change. Use when an icon "looks like a template", "means nothing", is inconsistent with its...

Draw the site's own graphics as SVG in the lab - an icon beyond the families, a mark beside copy, an illustration, an animated logo or brand mark, a short 2D loop - starting from the artwork the site already has, looked at on the site's tokens in both schemes, at the sizes it ships and enlarged, and on a phone (`pnpm kit lab serve`), checked frame by frame and as the embedded image, shipped pre-rendered as the file a page embeds or as inline Icon data (`pnpm kit lab render`, `pnpm kit icons a...

Turn a plain-language guardrail into a handrail Rule file, prove it matches, and validate it. Use when the user wants to add, write, or create a handrail rule, says "never let the agent ...", "block ...", "ask me before ...", "warn me when ...", or wants a Global rule relaxed in one repository.

Scan a new or unfamiliar repo and generate AGENTS.md with stack, commands, structure, and agent context.

The canonical home for how each kind of writing we produce gets written: code comments, review comments, PR descriptions, commit messages, chat messages, workflow artifacts, and tickets. Load the matching reference before producing one of these, when asked how something should be written, or when defining or polishing a style. Skills that produce output point here instead of restating rules.

Audit and adjust the writing a piece of work produced, including workflow artifacts, code comments in the diff, PR descriptions, and unpushed commit messages. Use once draft PRs are open and before they go ready for review, or when asked to audit comments, descriptions, or writing. Running it before the PRs exist leaves their descriptions unaudited.

Set up a repository with a person's standing preferences — configs, lint/format, tsconfig, CI, GitHub settings, AGENTS.md, IDE settings. Use when the user wants to "imprint" a repo, scaffold a new project to their conventions, apply their house style, or bootstrap tooling for a fresh or existing repo. Looks for a `<username>/imprint` repo and applies its IMPRINT.md. Pass `fork` to instead create the caller's own imprint repo from this one.

Remove Xaedalon Factory from this machine — stop its agents, take out the software, and report exactly what was left and why. Use when the user asks to uninstall, remove or get rid of Factory, or invokes /factory-uninstall. Not for removing one project or task from Factory (that is the board, or `factory project`).

Remove Xaedalon Factory from this machine — stop its agents, take out the software, and report exactly what was left and why. Use when the user asks to uninstall, remove or get rid of Factory, or invokes /factory-uninstall. Not for removing one project or task from Factory (that is the board, or `factory project`).

The author's aesthetic philosophy, prose style, and creative sensibility — the single source of truth every editorial skill judges against. Ships unconfigured; the first direct invocation runs a one-time interview that rewrites this file into the author's real style reference.

Use when an accepted PRD under docs/prd/ needs a technical architecture before specs are written (step 2 of orch-plan). Reads the code and the PRD, writes docs/arch/NNN-<slug>.md with components, contracts, decisions and the phase/package split orch-spec turns into tasks. Never writes code, specs or tasks.json.

Quality gate for a repo: one bun script behind git hooks (pre-commit, commit-msg, pre-push) and the Stop hooks of Claude Code, Codex and pi. Blocks test tampering, CRAP and complexity over the bar on changed functions, low diff coverage, new cycles, dead code, clones, secrets, vulnerable dependencies and AI attribution in commits. 12 languages, one .quality.toml per repo. Use when: quality gate, CRAP, complexity, ratchet, baseline, hotspots, dead code, import cycles, wire a repo into the gate...

Routes, designs, evaluates, and refines deterministic rules from concrete anti-patterns. Use when creating repository policy or a Python, TypeScript, Markdown, SQL, IaC, or config rule; checking whether an upstream rule already covers a problem; calibrating false positives on local corpora; or deciding whether a new rule is safe to introduce as a warning or error.

Shared read-only discovery, review, and reporting protocol for every Sarj judgment-layer audit command.

Task management patterns and best practices for the gur CLI tool

Test-driven development with red-green-refactor loop. Use when user wants to build features or fix bugs using TDD, mentions "red-green-refactor", wants integration tests, or asks for test-first development.

Grilling session that challenges an existing plan or design against the project's domain model, sharpens terminology, and updates documentation (CONTEXT.md, ADRs) inline as decisions crystallise. Use when a plan or design already exists and needs stress-testing; if no plan exists yet, use feature-planner first.

Upstream feature planning — requirement clarification, constraint discovery, design outline, vertical-slice task breakdown ordered by risk. Hands off to grill-with-docs and tdd. Use when starting new feature work from a vague request and no plan exists yet; if a plan or design already exists, use grill-with-docs to stress-test it instead.