jezweb

Australian business English for professional writing — warm, direct, EN-AU spelling (colour, organise, centre), no filler words. Use whenever the user is writing for an Australian audience: emails, chat messages, proposals, client communications, blog posts, web copy, or any business writing. Apply to drafting, editing, and tone-checking professional text.

Hit the Cloudflare REST API directly for operations that wrangler and MCP can't handle well. Bulk DNS, custom hostnames, email routing, cache purge, WAF rules, redirect rules, zone settings, Worker routes, D1 cross-database queries, R2 bulk operations, KV bulk read/write, Vectorize queries, Queues, and fleet-wide resource audits. Produces curl commands or scripts. Triggers: 'cloudflare api', 'bulk dns', 'custom hostname', 'email routing', 'cache purge', 'waf rule', 'd1 query', 'r2 bucket', 'kv bulk', 'vectorize query', 'audit resources', 'fleet operation'.

Generate custom favicons from logos, text, or brand colours. Produces favicon.svg, favicon.ico, apple-touch-icon.png, icon-192/512.png, and web manifest. Use whenever the user wants a favicon, mentions replacing a CMS default favicon, converting a logo into a favicon, creating branded initials icons, or troubleshooting favicon not displaying / iOS black square / missing manifest.

Run an independent code review using the OpenAI Codex CLI in headless mode. Gets a second opinion from a different model family (the current Codex models) on recent changes, a PR, a commit, or the whole app — covering bugs, regressions, security, data consistency, UX/state bugs, performance risks, and testing gaps. Saves a severity-prioritised report to .jez/reviews/. Triggers: 'codex review', 'review with codex', 'independent code review', 'what does codex think', 'get codex to review'.

Deep research and discovery before building something new. Explores local projects for reusable code, researches competitors, reads forums and reviews, analyses plugin ecosystems, investigates technical options, and produces a comprehensive research brief. Three depths: focused (30 min), wide (1-2 hours), deep (3-6 hours). Triggers: 'research this', 'discovery', 'explore the space', 'what should I build', 'competitive analysis', 'before I start building', 'research before coding'. Not for cited fact-checking research reports (a separate harness does those); this is pre-build product discovery.

Plan and execute entire application builds. Generates phased delivery roadmaps, then executes them autonomously — phase by phase, committing at milestones, deploying, testing, and continuing until done or stuck. Modes: plan (generate roadmap), start (begin executing), resume (continue from where you left off), status (show progress). Triggers: 'roadmap', 'start building', 'resume the build', 'keep going', 'build the whole thing', 'execute the roadmap', 'what phase are we on'.

Walk through a live web app AS a real user to find usability + behavioural bugs that static reviews miss. REQUIRES proof of interaction (typing, clicking, sending, observing) before any verdict — a sweep that didn't interact terminates with verdict 'Incomplete'. Walks threads, exercises every element, runs the multi-pane stress matrix, visual polish sweep, component perfection checklist, automated a11y (axe-core), pragmatic performance budget (LCP/CLS/INP), scenario battery (11 scenarios), and stress recipes including the real-flavour data battery. Hard gates: console errors/warnings = 0, network 5xx = 0, layout collapse = 0, axe Critical/Serious = 0, perf budget green. Audit-the-audit meta-check rejects rushed reports. Each finding has reproduction steps, evidence path, and suspected code location. Trigger with 'ux audit', 'walkthrough', 'qa sweep', 'audit the app', 'dogfood this', 'check all pages', 'find what's broken', 'stress the UI'.

Compare UX patterns across multiple reference apps using pattern libraries produced by ux-extract. Reads 2+ pattern-library.md files, walks them category by category, identifies where apps converge (strong signal), where they diverge (genuine design choice), what's unique to one app, and what's absent across the set. Produces an opinionated comparison document with recommendations for a new build. No browser needed — pure markdown analysis. Trigger with 'compare UX patterns', 'how do top apps handle X', 'ux comparison', 'pattern comparison across reference apps'.

Evaluate multiple choices side-by-side based on relevant criteria, then present a comparison the user can act on. Use when the user asks to compare options, choose between alternatives, or weigh up X vs Y.

Day-of-week rotating proactive sweep — Mon=connections / Tue=routines / Wed=skills / Thu=memories / Fri=findings / Sat=cross-cutting / Sun=weekly summary. Forces outward scan instead of inward thrash. Routine-only; load this skill into a daily routine that fires AssistantAgent. Adapted from goanna's caretaker.

Use when an agent has had N consecutive quiet runs with nothing to do. Pivots to bounded inward consolidation (promote findings, refresh memory, audit asks) anchored in own data — NOT generic news summarisation. Adapted from goanna's reverie cycle. Routine-only; loaded by the routine that detects the quiet condition.

Use to track open questions you owe the user (asks) and time-bounded commitments you owe yourself (tasks). Write OPEN entries when you commit; promote to CLOSED when answered/done. Survives session compaction. Adapted from goanna's asks.md/tasks.md pattern.

Reviewer criteria for AI-generated code via with_review. Checks for security issues (injection, auth bypass, secret leakage), correctness, and project conventions. Use as criteria.skill when generating code that touches data, auth, or external systems.

Reviewer criteria for outbound email drafts via with_review. Checks tone-recipient match, professionalism, no AI tells, no placeholder leakage. Use as criteria.skill argument when reviewing email drafts before send.

Reviewer criteria for AI-generated summaries via with_review. Checks faithfulness to source, no hallucinated facts, appropriate length, key points preserved. Use as criteria.skill when summarising articles, documents, transcripts, or batch-task outputs.

Default reviewer criteria for the with_review tool. Generic quality bar — accuracy, intent-match, no hallucinations, clarity. Use as the criteria.skill argument when calling with_review without a domain-specific reviewer skill.

Use before calling any visual or rendered output done, a UI, a layout, a generated image, a rendered document, a chart. Render the real thing and look at the actual result (playwright or Chrome screenshot, computer-use, or the project's own render/export command) before you ship it. Don't report "looks good" on something you never viewed.

Use when you want a lightweight, claimable task board that lives in the repo as markdown, for parallel multi-agent work or your own backlog, where a card is a file and the folder it sits in is its status. Covers the folder-board layout, claiming and completing by moving the file, worktree-safe concurrency for a fleet, frontmatter conventions, discoverability one-liners, and an optional zero-dependency local web UI to view and drag cards. Triggers, "kanban", "task board", "track parallel agent work", "claimable work queue", "what's in flight".

Use after any non-trivial build, before commit or deploy. Convene two or more frontier models from different families and providers to review the change for Critical/High/Medium/Low findings; fix cross-validated Criticals before commit and Highs before deploy. Also when the same bug won't fix on the first attempt, or for a whole-codebase pass before a release.

How to plan and approach a non-trivial build or feature with an agent, before writing phases or code. Covers the Phase-0-refactor catch, thinking in contracts, discovering the data structure from real instances before designing it, writing the proof of done first, verifying by inspection, splitting plan from spec, trusting skills over elaborate code, the moving code-vs-agent boundary, and matching a precise brief literally with a sign-off preview before shipping. Load when starting, scoping, or planning any substantial piece of work.

Use when deciding how to arrange agents for a piece of work: one agent or several, which pattern (single loop, pipeline, routing, parallel, orchestrator-workers, evaluator loop, triage funnel), how to match roles to model temperaments, how to bound a model-driven control loop so it's safe to trust instead of hand-coding a deterministic spine, and the mechanics that keep multi-agent work safe and cheap. The layer above agent-delegation (one-off hand-offs) and run-a-role-agent (one ongoing role).

Use when setting up an ongoing agent dedicated to a role rather than a one-off task, a webmaster, researcher, bookkeeper, support triager. Covers shaping the persona and its lane, populating a specialist with real reference (not a generic shell), running it with one bounded output per run, putting it on a heartbeat, and the observer disciplines for a role that watches a live human system and may ramp up to running it (observe without consuming the signal; the observer-to-actor trust ladder). For backlog-grinding work that tunes its own procedure, see run-a-self-refining-loop.

Use when writing or revising any prompt, a system prompt, an agent brief, an extraction or classification prompt, a sub-agent contract, or when a prompt-driven feature underperforms. Covers the three parts a prompt needs (goal, contract, knowledge), worked examples over templates, the directive vs goal-oriented style lever, matching model temperament to the task, right-or-null for derived facts, building minimal-first (start small, add one thing at a time, never write big and cut back), and pruning the bloat ratchet.

Use before shipping anything agents or newcomers will read cold: a README, a CLAUDE.md or workspace brief, a skill, a playbook, a template repo, onboarding docs, a production prompt or sub-agent contract. Two moves: audit every claim (derived / checkable / judgement), then run a clean-context dress rehearsal where a no-context agent actually uses the docs and reports where it guessed. Catches insider references, drifted counts, broken pointers, and stale facts the author's context papers over.

How the worker approaches deep work — research → plan → ship → verify → log. Use when the worker has been handed a task that requires more than a single tool call.

Graduate findings into wiki entries. Read findings/ in any building, judge which patterns are sturdy enough to promote, write entries via the wiki MCP. Use when the librarian is doing curation passes or when boss flags accumulated learnings for promotion.

How boss routes work across the town. The discipline of delegation — when to hand off, who to, with what brief, what to expect back. Use when boss is about to do work itself instead of delegating.

Pull structured knowledge from unstructured sources (emails, web pages, transcripts, files, conversations) into wiki-ready shape. Use when the librarian is asked to "extract what we know about X" or to file something from external systems.

How the scout scans environment, industry, tools, world. Surfaces patterns that inward-focused roles wouldn't notice. Use when boss asks "what's brewing?" or scout fires on its own initiative.

How to run, define, and report on Workflows — the standing jobs the cortex owns. A Workflow is a Goose recipe (the goal) + a trigger.yaml (when it fires + trust). Use when a workflow fires (a matching file lands in inbox/, a schedule, or a webhook job), when the owner says "put my X on a workflow" / "run my workflows", or to pause/change one.

Drafts the monthly investor update — pulls metrics, surfaces wins + struggles, formats per the founder's voice. Use on the 1st of each month, or when boss asks for a draft, or when investor-relations fires a scheduled monthly routine.

After filing a meeting, customer call, investor call, or any conversation entry, scan the body for explicit commitments (I'll ship X by Y, send Z by Friday) and create structured commitment entries in wiki/commitments/. Use whenever a new meeting note lands; also use to retroactively scan an existing entry for missed commitments.

Compute current runway, burn trend, and forecast end-of-runway. Use when boss or bookkeeper is asked for runway, on a scheduled monthly check, or before any fundraising conversation.

After a customer call (transcribed or hand-noted), file a structured summary to wiki/research/ and trigger extract-commitments. Use when founder finishes a call and says "@customer-success file this call" or similar.

Move a candidate through the hiring pipeline — sourced, contacted, first-call, case-study, offer, hired, passed. Captures notes per stage; never lets a candidate ghost-drift. Use when recruiter advances a candidate or the founder adds a new one.

Install the OfficeCLI binary (create/read/edit .docx/.xlsx/.pptx with no Microsoft Office) and register it as a Goose MCP server. Use when the user wants Office-document capabilities, asks to 'install officecli' / 'add Word/Excel/PowerPoint support', or a skill/app needs to generate Office files.

Draft an email newsletter — or brainstorm topics for one — from the business's own recent work and knowledge. Use when the owner says "write the newsletter", "what should this month's newsletter cover", "draft an email to the list about X".

Turn something worth sharing (a win, an announcement, a link, a blog post) into a platform-shaped social post — LinkedIn, X, Facebook, Instagram. Use when the owner says "post this", "share this on X", "write a LinkedIn about Y".

Install the official Google Workspace CLI (gws) so the agent can drive Gmail, Drive, Docs, Sheets, Calendar, Chat and Admin from the shell. Use when the user wants Google Workspace access on a box, asks to 'install gws' / 'add Google Workspace', or a workflow needs to read mail / write a Doc / update a Sheet. gws is a shell CLI, not an MCP server — the agent calls it through its developer/shell tool.

Turn a construction job folder (PDF plan sets, engineering, approvals, specs) into a queryable plan pack — synthesised sheet register, object database with provenance+confidence, takeoff quantities with shown workings, topic wiki, per-sheet renders. Use when asked to "build a plan pack", "index these plans", "do a takeoff", or process a folder of construction drawings for AI querying.

Turn a trade business's existing files (past plans, quotes, price data) into a calibrated planpack workspace, ending with a trust report that re-quotes their own past jobs blind. Use when onboarding a trades business to planpack, processing a back-catalogue of past jobs, or asked to "set up my business on planpack".

Audit a website's health with Magpie — overall score, Core Web Vitals, on-page SEO, security headers, robots/sitemap, and WordPress-specific checks. Use when the user wants to audit a site, check if their website is healthy, run an SEO check, understand why a site is slow, or asks "how is my website doing".

Track Google rankings and find competitor keyword gaps with Magpie Pro. Use when the user wants to track their rankings, see where they rank for a keyword, ask "am I ranking", find what keywords competitors rank for, or run a keyword-gap analysis.

Audit a domain's DNS records — MX, SPF, DKIM, DMARC, CAA, A/AAAA — for correctness and best-practice. Use when onboarding a new domain or when an email-deliverability or SSL issue surfaces.

Check SSL certificate expiry across managed properties; flag anything within 30 days; renew where the renewal is automatable (Let's Encrypt, Cloudflare-managed). Use weekly via cron.

Set up dotbackup for a user — point it at their own Cloudflare R2 bucket, create the encrypted restic repository, store the secrets in their macOS Keychain, and write the config the app reads. Use when a user is installing dotbackup, says "set up my backups", connects dotbackup to Cloudflare, or the app shows the "no config" setup screen. macOS, Phase 1 (backup).

The operating method for autonomous build-and-ship work. Load it when you're about to build, code, refactor, fix, review, or ship a web app or service, or when running a self-directed work loop, and you want to work the way the rest of the toolkit expects: the loop to run, which sibling skill to invoke, how they chain, and the disciplines that keep the work honest. Not a router (the harness already picks skills) and not a task tracker — the method an agent runs WHILE doing the work. Triggers with 'shipwright', 'how should I work on this', 'start building', 'work loop', 'ship this', 'autonomous dev loop', 'what's the right way to build this'.

Audit a codebase so every comment, docstring, and doc file is a TRUE and useful reflection of the code — for the next AI agent (or human) who reads it before changing anything. Finds stale, inaccurate, misleading, and missing comments (and stale README/CLAUDE.md/ARCHITECTURE/config notes), produces structured findings, and splits them into a confidence/complexity/blast-radius-gated auto-fix set vs a human-review queue. Module-by-module, high-signal, never a comment-stripper. Use when the user wants to verify comments match the code, clean up stale/misleading comments or docs, make a codebase trustworthy for AI agents, or asks to "run honesty". Triggers: honesty, comment audit, do the comments match the code, are the comments accurate, stale comments, trustworthy comments, audit the docs against the code, comment truth.