ai-registry

Onboard any codebase into the bp product blueprint. Checks which artifacts the repo already has (context map, app-structure v3 map, events, lens-scan store, registration) and produces only what is missing: a v3 map with domains, capabilities, features and evidence-based stages; an events log from git history; static metrics in the lens-scan store; LLM judgement facets for every lens a judge can…

1 skill

Create or refresh a repo's AGENTS.md so an agent joining the codebase gets commands, architecture and constraints without guessing. Use on a repo with no agent guidance, or when the existing file has gone stale.

1 skill

Heavy structural codebase scan - surfaces weak patterns to upgrade and strong patterns to codify, and proves whether the repo's own gates (lint rules, hooks, CI checks, type rules, test guards, ratchets) still fire, producing ADR-style decisions with a durable cross-session backlog. The highest-risk / highest-payoff scan in the suite; walks the context map area by area. Modes: scan (one theme…

1 skill

Give a project its first real CI gate, built only from commands the repo already declares, and ratcheted so it goes green on day one. Use when a repo has no CI, has CI that has never passed, or has CI nobody trusts.

1 skill

Run the exact checks CI enforces before you push, so a red pipeline is never how you find out. Use before every push, before opening a PR, and after an agent finishes a batch of edits.

1 skill

Turn a red build into a located first cause and a scoped fix proposal, without scrolling the whole log or weakening the check. Use when CI fails and you need the actual reason.

1 skill

Spend the Claude Code cloud-session credit as a parallel engine beside local agents. Ships a self-contained brief to a `claude --cloud` session that clones the GitHub origin, works on branch claude/cloud-<id> and returns one pull request with a RESULT.md for local review. Two ways in: explicit dispatch (a plan by default, --go to launch) and overflow - fan-out skills queue cloud_ok briefs before…

1 skill

Evaluate this repository against the registry standards that govern it, one context at a time, and keep the verdicts. Reads .ai/registry-map.json (the generated join between this repo's contexts and the registry's subjects), picks the highest-value unevaluated or stale pairs, reads the governing golden path and techniques against the context's real code, and writes back conformant / deviation /…

1 skill

Consult the ai-registry's knowledge bundles before a product, architecture or domain decision: resolve the registry (local checkout or GitHub), pick the bundle(s) this repo consumes, match the task against subjects and techniques by their use_when triggers, read the golden path + the techniques that apply, apply them, and log the consult so the signals lane can count demand. Use before designing…

1 skill

Blind design contest between CLI agent seats (Claude Code, Codex CLI, Grok CLI, Antigravity agy, Muse). Each participant you name - engine:model@effort - builds three genuinely different prototype variants of one idea in its own workspace; a cross-family panel scores every variant blind on seven dimensions (wow, clarity at scale, wayfinding, interaction, craft, concept, utility); the host adds a…

1 skill

Strict, evidence-first triage of ONE already-built feature or ONE architecture redesign, ending at a human gate the method itself may never pass. Bounded members each answer a single question - worth to the user, workmanship, prior art, resilience, running cost, undo cost - reading only an evidence pack built from the tree, never the implementer's own account and never another member's answer.…

1 skill

Use when a prototype or app already exists and the job is to make its user experience good, or to finish it. Walks the product's user journeys screen by screen with a CX/UX lens: at each stop a short overview of the screen as the user meets it, 3-6 specific improvements graded by impact and effort, each naming its heuristic and the project's design doc, the user's own read and expectation for…

1 skill

Spend a ChatGPT/Codex plan on daily fleet hardening under Claude review. Use for a recurring stabilize or optimize pass across the connected projects, when Codex tokens are plentiful and Codex builders alone are not trusted. After one handshake with the operator (projects, strategy, hours, RAM, push consent, focus), Codex gpt-6.1-sol at medium effort surveys each context read-only; the Director…

1 skill

Make a voice-first demo film or product video of a working app, compiled by Remotion from one film script - real app footage on one lane, motion graphics (a typed plate library plus 1-3 bespoke hero compositions) on another, ElevenLabs narration measured before anything is recorded, and cue-spotted ElevenLabs music beneath. Use when asked for a demo film, product video, launch or explainer video…

1 skill

Autonomous infinite training loop for media-generation craft: each cycle reflects prior human verdicts into the consuming repo's prompt surfaces, scans the registry for the weakest knowledge dimension, researches one candidate technique, generates seed-matched A/B pairs with the LOCAL image/video stack, machine-judges them pairwise, and parks the cycle for human gating in the app's Foundry ->…

1 skill

Short, interactive, joins-first probe sweep of one codebase area. Picks a (context, lens) pair that neither explorer nor scan-sweep has run yet (scripts/pick.mjs reads the context map, explorer's coverage ledger and scan-sweep's history), runs a catalog of one-command cross-layer probes before sampling any file (a gate matrix over a route family, the repo's own idiom where it was not applied, a…

1 skill

Quarantine an intermittently failing test as tracked debt - owner, cause, expiry - instead of deleting it or leaving it to rot. Use before muting or skipping any test, and on a schedule to review the register.

1 skill

Use when the operator wants steady product momentum in ONE area of a codebase: an endless single-area companion loop that scans the area, proposes 5 development directions seeded from evidence (the area's registry-map deviation pairs and their golden-path techniques, the directions ledger, sibling passes and backlogs, and which earlier friend branches were merged), executes the picked one in a…

1 skill

Time-boxed, owner-judged hackathon between CLI agent seats (Claude Code, Codex, Grok, Antigravity agy, Muse) on a BROAD track in a real repository. Use when the owner knows the area but not the feature - 'make audio usable for non-musicians' rather than 'build this panel' - and wants several minds to find the projects and race to build them against a wall clock. The host frames the track with the…

1 skill

Copywriting-grade, context-aware localization for any managed app - a transcreation loop (draft -> typed MQM estimate -> gated refine) with an engineering guardrail. Portable across i18n stacks via a per-repo contract artifact; keeps a glossary, per-locale style guides, and gold exemplar pairs in the target repo so run N+1 stays consistent with run N. Invoke with /i18n-translate <mode> [locale]…

1 skill

Take an existing web component that explains a concept with conventional web design (a card grid, a bulleted feature list, a static diagram, decorative art) and give it an illustration layer that carries the claim: read what the section is trying to prove, think divergently about how to abstract it graphically, then prototype three directional illustration variants inside the component behind a…

1 skill

Revitalize and unify the design of a web app: measure style divergence per module, gate a foundation (type scale, muting, colour roles) with the owner, stop new slips with ratchets and agent law, run a contest for a COMPOSITION KIT (the shared building blocks one level above buttons), port the finalists onto one real page so the owner chooses on the product, promote the winner at 0 measured…

1 skill

KPI Simulation - measure a project's KPIs locally, simulate user behavior with UAT-style Characters, and predict real-world targets from web benchmarks, writing a result.json the Personas app ingests into its KPI module (env-tagged simulated series + proposal-gated adjustments). Engine doctrine mirrors src/features/teams/sub_kpis/kpiSimPrompt.ts - the app dispatches the same contract into managed…

1 skill

Measure a product feature by feature through one lens or all of them in one read - development, security, design, or any lens whose manifest declares a scan block (judgement, or composed script / conform / judgement steps) - and record hard numbers over time in a local store inside the repo, then propose backlog items scored by the shared scan-finding/1 contract that each name the metric they…

1 skill

Generate images with OpenAI GPT Image 2.5 (Sunburst for detail, Flare for speed) or Leonardo AI (fallback), remove backgrounds, analyze with Gemini vision, and write SVG. For brand assets, UI illustrations, backgrounds, and icons.

1 skill

Compare models head-to-head on ONE declared LLM use case of a connected project, and return a rating table a human can act on. Picks the project and the use case from tracklight's use-case registry (registering them if the project has none), fans the identical task across an arm per model - hosted Nebius open-weight models, an OpenAI-compatible endpoint, and the local Claude Code CLI - mirrors…

1 skill

Run resumable maintenance after rapid development: reduce demonstrated code debt through small verified refactors, defect repairs, tests, dependency upkeep, and documentation corrections. Use for sustained codebase maintenance; product ideation and redesign belong to other workflows.

1 skill

Upgrade a generic UI icon or loading/empty state into a traced, motion-animated SVG. Generates flat trace-friendly art (via /leonardo tools), validates with Qwen vision, vectorizes to a clean multi-path SVG, and emits a motion reveal component driven by the shared motion-preset library (draw, staggered-draw, fade-pop, float, pulse, hover-response, success-settle). For icon + loading-state visual…

1 skill

Launch-readiness orchestrator for taking a dev project (typically NextJS) from "advanced WIP" to first market release. Assesses 21 checklist items across 7 chronological phases (Define -> Rails -> Automate -> Harden -> Polish -> Market -> Launch gate), reports an honest scorecard, then walks each phase with batched select decisions and executes accepted work with subagents. Delegates infra to…

1 skill

Write, check and review native-quality ENGLISH web copy where it lives: i18n JSON catalogs, TS dictionary modules, hardcoded JSX text, MDX. Records the product's declared English (US/UK spelling, dash system, quotes, case per element) in a copy contract counted against the real strings, gates changed copy with a zero-dependency checker whose every finding cites a registry EN-* rule ID (baseline…

1 skill

Fetch npm package updates, analyze new features, and identify improvement opportunities for this app. Use when the user wants to explore what's new in dependencies or plan next development directions.

1 skill

Take a fresh clone of an app repo to a running, honestly-labelled install in one conversation: probe runtime deps, ask which connector capabilities the operator wants, collect keys into .env.local without ever echoing a secret, verify by booting the app, and hand back a capability matrix with three honest states per group (works / degraded with a stated fallback / hidden). Runs on generic…

1 skill

Session-after-session product perfection loop. The strongest available model at xhigh reasoning (currently Sonnet 5.5) directs - it walks the repo's context map context-by-context, proposes up to 5 challenged, high-value directions per context (features, design elevations, significant optimizations), gates them with the user until the pool is full, then orchestrates Sonnet-class builder subagents…

1 skill

Populate a newly managed repository with the data Personas needs to maintain and develop it - a context map, a feature (use-case) inventory, a triaged KPI set, and optionally simulated KPI data for a product that has not shipped yet. Contexts and features are assigned autonomously; KPIs are negotiated with the operator wave by wave. Scopeable - run all four lanes or just the ones you name.…

1 skill

Promote a specific, already-existing pattern in this repo into the Personas workspace knowledge library - the grounded CLI equivalent of the "New practice" button in Dev Tools -> Workspaces -> Knowledge Library. Unlike /practice-harvest (a broad <=15-item sweep the agent chooses), /promote is targeted at ONE pattern the operator names, reads the real code, and writes the evidence. Emits the same…

1 skill

Extract actionable improvements for a project from external sources (video, blog, article, raw text). Scores ideas against the codebase, buckets into Code / Template / Credential, and persists findings to an Obsidian memory vault.

1 skill

Long-running quality sweep that walks a repository context by context, reads each area's code once, judges it through every scan lens, and lands what it can PROVE itself with atomic commits. With no arguments it runs the STABILIZE loop - bug hunting, UI perfection, performance - picking the least-covered context each round and keeping a per-context lens-coverage ledger so a codebase gets swept…

1 skill

Milestone-driven ship-readiness loop for any app repo, resumable across sessions. Boots by profiling the stack, running the verification gate once, and fanning out read-only audit lenses into a 9-dimension scorecard plus an append-only numbered backlog; when the repo's overlay declares value journeys, a value ledger (one readiness light per journey) sits above the scorecard, every item carries a…

1 skill

Turn a vague product idea (a "sparkle") into a complete, grounded design through waves of select/multi-select questions - then orchestrate the build. Targets exactly which contexts/files the idea touches, scouts them before asking anything, converges the design across four perspectives (functional, UX, UI, performance/architecture), and executes via builder subagents in a worktree under Director…

1 skill

Drain the fleet's version debt against the registry in one sitting. Run from the registry checkout when a bundle landing, a context-map rebuild, a rename or an uncommitted converge rebuild has left registry maps behind: classify every checked-out project's .ai/registry-map.json (modified but equal to a rebuild from HEAD: generator-dirty, committed; verdicts differ: somebody's work, never…

1 skill

Take over an unknown or vibecoded fullstack app and bring it under control toward production readiness: a phased, gated sequencer over the registry's existing skills. Recon the stack, boot it, map it, score readiness, scan it through the built-in lenses plus four takeover lenses (API and LLM call hygiene, security depth, backend performance, money and privacy) with an Opus judge, import the Azure…

1 skill

Prove a change works before it is committed by writing or extending a test that fails first and then passes. Use whenever you change behaviour, fix a bug, or accept AI-generated code.

1 skill

Hunt a codebase's most valuable feature area and evolve it with a cross-vendor council. Use when you want the product's strongest part made markedly more valuable - not swept for defects - and a person to choose what gets built. tiger ranks feature areas by value evidence (context-map groups and the use cases they serve, git heat, product docs, the bp product tree), writes a territory dossier for…

1 skill

Simulated User Acceptance Testing driven by Characters (representative users with jobs-to-be-done), not feature/code coverage. A capable LLM verifies each user journey in two chronological certification levels - L1 theoretical (over a code-derived surface model, cheap + mass-parallel) then L2 empirical (real browser against the live app, serial) - plus, for a journey whose core is a conversation…

1 skill

Model a product's value per user journey from its UAT Characters - time and money saved against the LLM-less way, risk avoided (mis-hires, wrong decisions, compliance exposure), and segment gates - normalized to a reference journey so the output ranks work by user value instead of by what a code scan found. Sits UPSTREAM of code-derived KPIs - a KPI worth adopting is one derived from a journey's…

1 skill