Every pack is a line. Every skill is a stop.

1. Departs from your brief

   “Set up our brand charter, then audit our landing page against it.”

2. 01Premium Design Foundations
3. 02Optical Typography Craft
4. 03Spatial Composition Discipline
5. 04Premium Color Discipline
6. 05Material and Depth Language
7. 06Premium Motion Feel
8. Express past 47 more stops ↓
9. 54Pre-Ship Creative Gate

10. End of the line

    Every stop on this line, with a preview, is on the pack page. See it →

1. Departs from your brief

   “Write the asset contract for our game's characters.”

2. 01Game Asset Contract
3. 02Blender Headless Harness
4. 03Procedural Creature Sculpt
5. 04Body Variant Frames
6. 05Game Rig and Skin
7. 06Bake Atlas Textures
8. Express past 2 more stops ↓
9. 09Asset Review Gates

10. End of the line

    Every stop on this line, with a preview, is on the pack page. See it →

Gate results are from the Ledgerline sample run (Ledgerline is a fictional company).

1. Departs from your brief

   “Run a production readiness review on this repo.”

2. 01Readiness ContractPASS

   Detects the stack and tools, then writes and validates readiness-contract.json: tier, SLO, release base, staging host, backups and waivers.

3. 02Dependency and Secret ScanFAIL · 4 blockers, 1 major

   Every lockfile parsed and looked up in OSV with severity and the smallest safe upgrade, plus live credentials and tracked .env files found in the working tree and git history (it reports exactly how many commits it read; shallow or truncated history fails the gate), with evidence redacted.

4. 03Migration SafetyFAIL · 3 blockers, 4 major

   Lock, rewrite and compatibility findings for pending Postgres migrations, then a rehearsal on a throwaway Postgres with seed rows: locks per file, measured table rewrites and a down/up round trip.

5. 04Rollback PathFAIL · 2 blockers, 3 major

   Whether the previous release can be redeployed onto the new schema, PITR and restore-test status, and a drafted rollback runbook with commands.

6. 05Observability CoverageFAIL · 6 major

   A route and job inventory with missing error tracking or error middleware, swallowed exceptions, outbound calls without timeouts, health endpoints and unmonitored jobs.

7. 06Config DriftFAIL · 2 blockers

   Env vars the code reads compared with .env.example and each deployed environment: secrets behind public prefixes, staging/production drift and stale variables.

8. 07Staging Load TestFAIL · 2 major

   An open-loop, fixed-rate load test against an allowlisted staging host, judged against your SLO.

9. 08Readiness Report

   One evidence-bound decision (GO, GO with accepted risks, or NO-GO) in Markdown and HTML, refusing missing or stale gates.

10. Arrives · sample run on Ledgerline fictional

    • Launch candidate: NO-GO, with all 29 planted findings raised
    • Fixed twin: GO, with every gate PASS (3 minor notes)

    See the full run →

Gate results are from the Quorvel sample run (Quorvel is a fictional company).

1. Departs from your brief

   “Set up our SaaS metrics from our Stripe and bank exports.”

2. 01Finance Exports Intake

   Every export placed by its exact headers (files it cannot place reported), each bank account's exact columns, date format and failed-row filter proposed, and every Stripe price id mapped, in a first contract.

3. 02Finance ContractPASS

   A validated finance-contract.json: price kinds and intervals, MRR policy, amount units, FX source, statement balances, rules, contracts invoiced outside Stripe, payroll split, plan and waivers, with bank headers, dates and every price id checked.

4. 03Stripe Revenue IngestFAIL · 3 blockers, 7 major

   Every Stripe invoice line normalized to monthly value in your base currency (annual plans, discounts, prorations, trials, setup fees, refunds, disputes, FX at your rates checked against Stripe), every invoice reconciled to its lines to the cent, no month or invoice missing, and every dropped row listed with its reason.

5. 04Bank Cash IngestFAIL · 2 blockers, 6 major

   Bank statements proved row by row against their running and closing balances, every row categorized by your rules, transfers netted, every Stripe payout matched, receipts from customers outside Stripe matched to their contracts, and monthly cash collected, gross and net burn.

6. 05MRR Movements and CohortsFAIL · 3 major

   The monthly MRR waterfall (new, expansion, contraction, churn, reactivation, FX), ARR, logo churn, trailing-12 NRR and GRR and cohort tables, rolling forward exactly and tied to the subscriptions export.

7. 06SaaS Unit EconomicsFAIL · 3 major

   CAC, ARPA, gross margin, CAC payback and burn multiple over a trailing window, each input cited to bank rows, and BLOCKED rather than estimated when an input is missing.

8. Express past 3 more stops ↓
9. 10Investor and Board UpdateFAIL · 6 blockers, 1 major

   The board update in Markdown and HTML with a source on every number, accepted risks and founder-stated facts labelled, refused while any gate is missing, stale or failing.

10. Arrives · sample run on Quorvel fictional

    • Month-end close as first exported: Every planted trap raised; the board update refused to write
    • Fixed twin: Every gate PASS, the update written, and every metric equal to the ground truth to the cent

    See the full run →

Gate results are from the Parcelwise sample run (Parcelwise is a fictional company).

1. Departs from your brief

   “Set up evals for our LLM app from our production traces.”

2. 01Eval ContractPASS

   eval-contract.json with adapters for baseline, candidate and judge (a command or allowlisted endpoint, keys stay in your env), slices, graders, thresholds, budgets and waivers, validated; example Anthropic and OpenAI adapters.

3. 02Trace Eval SetFAIL · 5 major

   A frozen dev/test eval set from your JSONL, CSV, chat-log or OpenTelemetry traces: PII scrubbed before storage, duplicates removed, prompt-example copies excluded, stratified by slice, with a hash manifest that catches drift and leakage.

4. 03Eval Case LabelsFAIL · 6 major

   Confirmed reference labels and pass/fail verdicts from your people, with two-labeler agreement (Cohen's kappa) and adjudication, plus the runner that calls your app.

5. 04Eval GradersFAIL · 1 major

   Deterministic graders (exact, regex, JSON Schema, numeric, required facts, citations in context, rubrics) and an LLM-judge grader, each tested against known-good and empty answers.

6. 05Judge CalibrationFAIL · 3 major

   Judge versus human agreement, kappa, confusion matrix, length bias and position bias; an uncalibrated judge never counts.

7. 06Eval Regression GateFAIL · 24 major

   Baseline versus candidate on the frozen set with repeats for flaky cases, McNemar's exact test with Holm correction overall, per dimension and per slice, an honest power statement, and a pinned GitHub Actions workflow that can ship from a fresh checkout.

8. Express past 2 more stops ↓
9. 09Evals Release Report

   One evidence-bound decision (SHIP, SHIP WITH ACCEPTED RISKS, or HOLD) in Markdown and HTML, refusing missing or stale gates.

10. Arrives · sample run on Parcelwise fictional

    • Release candidate: HOLD, with every planted finding raised and every planted count matched
    • Fixed twin: SHIP, with every gate PASS (2 minor notes)

    See the full run →

Gate results are from the Quaywise sample run (Quaywise is a fictional company).

1. Departs from your brief

   “Run a technical SEO audit of our site.”

2. 01SEO ContractPASS

   Writes and validates .seo/seo-contract.json: production origin, the mirror to crawl, allowed hosts, crawl limits and rate, page templates with expected schema, priority pages, locales, keyword exports and waivers.

3. 02Site Crawl AuditFAIL · 1 blocker, 17 major

   A polite crawl that obeys robots.txt for its own user agent, retries failing requests and samples every template and locale, with findings per URL: redirect chains and loops, broken links, canonical and sitemap problems, soft 404s and host variants, robots rules blocking CSS or JS, titles, H1s and Open Graph, hreflang reciprocity and untranslated fallbacks, mixed content, near-duplicate and thin pages.

4. 03Structured Data CheckFAIL · 3 major

   Every JSON-LD block checked against the properties on Google's Search Central pages (each rule cites its page) and against the visible page (prices, ratings, FAQ text, headlines), plus the schema each template should carry.

5. 04Internal Link GraphFAIL · 3 major

   Click depth with the shortest path, orphan pages, sitemap coverage, internal PageRank and anchors, and priority pages that sit too deep.

6. 05Core Web Vitals LabFAIL · 3 major

   LCP, CLS and TBT from Lighthouse for one page per template with priority pages first, the median of three runs, with optional real-user field data from your own Chrome UX Report key; lab and field labelled, unreliable runs flagged.

7. 06Keyword Cluster MapFAIL · 2 major

   Your Search Console and keyword-tool exports clustered by shared ranking pages, each cluster mapped to its page, with cannibalisation, unindexable ranking pages and gaps.

8. Express past 2 more stops ↓
9. 09SEO Re-crawl Report

   A re-crawl diff over what both crawls observed (fixed, still open, regressed, unverified) and one evidence-bound audit report and content plan in Markdown and HTML that refuses missing or stale gates.

10. Arrives · sample run on Quaywise fictional

    • Audit crawl: FAIL, with all 37 expected findings raised and every keyword clustered as in the ground truth
    • Re-crawl after fixes: FAIL: 47 findings fixed, 1 regression introduced by the fixes, 1 major still open, none unverified (94 URLs observed in both crawls)
    • Fixed twin: PASS, with every gate PASS on a crawl and a re-crawl

    See the full run →