Fast orientation map for a zero-context session in the Narraitor repo. Use FIRST when starting work here, when asked "where does X live", "how is this repo organized", "what's the stack", "where do I look for the stores / AI code / themes / tests", or to check WHICH docs are trustworthy before acting on one. (Asking…
Isolation-vs-integration parity reasoning and the component promotion ladder for Narraitor. Use when something "works in Storybook" but not the app (or vice versa), before calling any component done/integrated, when writing stories with withStores or MSW, or when deciding whether Storybook-green is enough evidence.…
What counts as evidence in Narraitor - the test tiers (unit, Storybook, visual regression, e2e, manual QA walkthrough), what each can and cannot prove, acceptance thresholds, how to add tests, and how to read failures. Use when deciding which tests a change needs, interpreting a red or suspiciously green run, planning…
Port inline styles from a reference/demo component to production CSS in the Narraitor design system. Rigid 6-phase process that diffs inline styles against production stylesheets, ports gaps using design tokens, and removes inline styles from production components. Use when porting design system styles from any…
Analyzes a single GitHub issue and produces a detailed technical specification ready for implementation. Fetches the issue, discovers existing patterns in the codebase, defines scope boundaries, and produces an implementation plan with MVP tests. Use before starting work on an issue to understand it fully. Trigger on…
Inspect and debug live web pages using the Chrome DevTools Protocol via the bdg CLI. Use for targeted DOM queries, computed style checks, layout debugging, and CDP method calls without loading a full page snapshot. Prefer over chrome-devtools-mcp for token efficiency. Trigger on: "inspect element", "check computed…
Automatically invoke when CI checks fail on a PR. Triggers on: any mention of CI failure, red checks, failed GitHub Actions, or "CI is broken/failing". Reads session handoffs and a known-issues log to avoid repeating failed fix attempts. Do NOT wait for user to ask — invoke as soon as a CI failure is confirmed. Not…
Kicks off a recurring code-health audit — the "embarrassing code" / overengineering sweep that turns vague tech-debt anxiety into a short list of measurable, validated simplifications to ship. Bakes in the pre-flight discipline (sync the correct base branch, confirm the active worktree and its dev server, list files…
Finds and removes unnecessary code comments - the ones that restate the code, cite an issue number, narrate refactor history, or wrap a one-line function in six lines of JSDoc. Sorts every find into auto-fix, propose, or never-touch, so the WHY comments survive the sweep. Trigger on: "comment reaper", "reap comments"…
Choose Your Own Adventure mode. Use when the user invokes /cyoa, asks for "cyoa mode", "choose your own adventure", or asks to make a development task feel like a game. Frames a real software task as a branching adventure where the user picks design decisions at every meaningful step. Produces real, working code — the…
Identifies and removes dead/orphaned code, tests, and Storybook stories — including trivial tests that check only obvious behavior or rare edge cases with no real value. Always validates that code is truly unused before deleting anything. Trigger on: "remove dead code", "clean up unused files", "delete orphaned…
Sniffs out stale and deprecated documentation — markdown files, code comments, docstrings, and commented-out code — then updates or removes it. Grounds every staleness call against the actual code before acting: auto-fixes the high-confidence rot, surfaces judgment calls for review. Trigger on: "find stale docs"…
Use when explaining code, debugging, or discussing architecture with Jack — map programming concepts to Drupal equivalents he already knows from years of Drupal/theming work. Triggers on code reviews, debugging walkthroughs, architecture discussions, inline code explanations, or whenever a non-Drupal framework concept…
Automatically run after any edit to a .css, .tsx, .ts, .jsx, or .js file that contains styling. Triggers on: new CSS rules, className changes, style props, token usages, or any visual/layout change. Checks the diff against the project's design system token inventory and living style guide (canon, if one exists). Flags…
An honesty gate that makes an AI back up its claims with real evidence instead of presenting guesses as fact. Use as an always-on background behavior, and invoke explicitly when the user says show your work, prove it, how do you know that, evidence check, what's your source, back that up, or when the AI is…
Generates GitHub-compatible markdown for screenshots using raw.githubusercontent.com URLs via the generate-github-image-markdown.sh script. Trigger on: "add a screenshot to this PR", "embed image in issue", "attach screenshot", "paste image into GitHub", "add screenshot to docs", "insert image in markdown", "show a…
Periodic and on-demand maintenance pass over a repo's open GitHub issue backlog. Re-labels miscategorized or unlabeled issues (bug/enhancement/question/invalid/duplicate) and bug lifecycle labels (needs-repro/needs-info) following the same decision logic as Anthropic's own dogfooded triage-issue command; re-scores…
Apply KISS principles to a diff or set of changes. Trims the change to the minimum surface required to achieve the stated goal — removes refactors, cleanup, style fixes, and "while I'm here" additions that aren't strictly necessary. Use when a diff has crept beyond its original scope or when you want to commit the…
Sweeps a codebase's text — comments, docstrings, markdown docs, and user-facing UI copy — for jargon that could be plainer, AI writing tells/tropes, and verbose text that could be tightened. Auto-fixes unambiguous swaps, surfaces judgment calls for review. Trigger on: "jargon audit", "plain language audit", "find…
Runs the post-merge workflow after a PR is merged. Phase 0 syncs the correct base branch and deletes the merged branch and its worktree so later work never starts from stale state. Phase 1 automatically closes the loop on linked GitHub issues (posts a completion comment, marks acceptance criteria done, closes the…
Autonomous PR review that also applies fixes for non-controversial issues — combines review, fix, test, and commit into a single pipeline. Unlike a read-only review, this writes code. Only use on feature branches where you're comfortable with automatic edits. Trigger on: "review and fix PR", "auto-fix the PR", "clean…