shipwright-devices
claudePinned mobile-mcp MCP server so the designer and the manual tester can drive local iOS simulators and Android emulators (list, tap, type, read the screen).
1 skill · 1 MCP server
Planner-led delivery workflow: requirements to traceable tasks, static-first design hand-over, backend/integration development, manual verification with evidence, token-cost tracking, guard hooks, and optional project-level wiring for Hindsight memory and context-mode.
Plugin shipwright-delivery from marketplace shipwright in vebg/shipwright. Commands read from the repo's manifest — 19 days ago.
This plugin runs code on your machine — its hooks fire automatically on events like tool use. The exact commands are listed below; read them before installing.
balanceCost balance of the workspace — every task with its status, group (feature or priority band) and USD cost, a header with the total spent and statistics (by status, average/median per done task, most expensive, forecast for the remaining tasks) and the budget lines from shipwright.json when set. Use
designStart the frontend-designer on the ready design parts (static-first UI) after the owner's go, then inspect its screenshots and evidence and close or return each design part. Use when ready tasks name the frontend-designer.
developStart the fullstack-developer on the next eligible tasks after the owner's go — returned tasks (verification failed) first, then ready backend and integration tasks — and record the hand-over. Use when ready implementation tasks exist.
doctorCheck the machine and the workspace before agents run — Python, Git, Node, Claude Code, the kit files and validator, the code repositories, the Hindsight runtime/server/hooks, context-mode, guard hooks, ignore lines, and the mobile/evidence tools (Xcode, adb, Flutter, .NET, ImageMagick, ffmpeg). Use
ownerThe owner's checklist — everything the owner must do that agents never do: versions to release since the last deployed one, environment/configuration lines, migrations, store steps, "For the owner" sections from evidence, open questions with their defaults, and HANDOFF's "waiting on the owner" items
planRun a planning pass as the planner — ingest new or changed requirement sources, update the spec, scope, architecture, decisions, diagrams, traceability and the dependency-ordered backlog, record open questions and validate the index. Use after requirements arrive or change, or when the owner asks to
questionsThe open-questions table for the owner — every question in docs/QUESTIONS.md §1 and every awaited input in §3 with its default in force, what it blocks and a one-line way to answer. Use when the owner asks "what do you need from me?" or before a planning session.
releaseSet up or run the kit's release bookkeeping — one shared product version tagged in every code repository per verified task, a release note per version, compatibility rules after the first public release. Use when the owner wants versioning/tagging or a release note, or when the planner needs to reco
resumeResume a Shipwright workspace — read the handoff and the backlog, recheck interrupted work, and propose the next batch (returned tasks first, then design parts awaiting inspection, then hand-overs awaiting the tester, then the next ready tasks). Use at the start of a session in an existing workspace
startGuided entry point of the Shipwright. Runs the discovery interview (product, users, surfaces, inspiration links/videos/apps, technologies already decided, constraints, brand, working agreement), collects the few decisions the kit needs, scaffolds the workspace (optionally with Hindsight memory and c
statusReport the state of a Shipwright workspace — backlog counts by status and verification, next ready tasks, open questions, blockers, Git status per repository and the token cost so far. Use when the owner asks where things stand.
verifyStart the manual-tester on the tasks handed over with verification pending, then read its evidence, replan follow-ups and update the handoff. Use after a developer hand-over.
workflowThe Shipwright's shared delivery protocol — task states and index fields, the manual verification flow, the static-first UI work split, design and screenshot workflow, engineering discipline, multi-repository rules and session sizing. Load when an agent needs the exact rule for a hand-over, a status
frontend-designerFront-end design owner for every surface (web pages and hosts, native mobile and tablet apps). Delivers the static presentation only — markup, styles and theming, screen/widget/theme files, motion, every UI state as a Development-only design route with hardcoded sample values, UI strings, assets, de
fullstack-developerBackend owner (API, realtime, persistence, auth, contracts, clients, build and release preparation) and the whole front-end integration that connects frontend-designer's static UI to real data; fixes every returned task, including designer-built UI, without redesigning it. Self-checks builds and run
manual-testerManually verifies tasks handed over by fullstack-developer (verification pending) by running the real API, apps and browsers against each acceptance criterion, including negative, authorization and concurrency checks; records evidence with real screenshots and sets the task done or returns it with d
plannerMain-session lead of the Shipwright. Turns requirements into traceable scope, architecture, decisions and dependency-ordered tasks; orchestrates frontend-designer, fullstack-developer and manual-tester one at a time; closes static design parts by inspection; keeps docs/HANDOFF.md current. Runs as th
store-compliance-reviewerOptional pre-submission review of mobile app builds and store listings against Apple's App Store Review Guidelines and Google Play's Developer Program Policy, using the project's vetted offline analyzers plus its own rule checklist; keeps docs/store-compliance/DASHBOARD.md with a GO / GO WITH NOTES
python3 "${CLAUDE_PLUGIN_ROOT}/scripts/session_start.py"
python3 "${CLAUDE_PLUGIN_ROOT}/scripts/skill_guard.py"
python3 "${CLAUDE_PLUGIN_ROOT}/scripts/agent_guard.py"
python3 "${CLAUDE_PLUGIN_ROOT}/scripts/policy_guard.py"
python3 "${CLAUDE_PLUGIN_ROOT}/scripts/token_usage.py" --quiet || true
python3 "${CLAUDE_PLUGIN_ROOT}/scripts/token_usage.py" --quiet || true
Pinned mobile-mcp MCP server so the designer and the manual tester can drive local iOS simulators and Android emulators (list, tap, type, read the screen).
1 skill · 1 MCP server
Matt Pocock's agent skills for real engineering (grilling, spec/ticket flows, TDD, code review, domain modelling and more).
39 skills
Core skills library for Claude Code: TDD, debugging, collaboration patterns, and proven techniques
15 skills · hooks
Talk like caveman. Cut filler. Keep technical details explicit.
22 skills · 1 command · 3 agents · hooks
Persistent memory for Claude Code: the next session starts where the last one left off.
29 skills · 1 MCP server · hooks
tasteskill 2.5.1: professional company websites at the figma.com and x.ai level, with one section per screen, real photographs, real product UI, effects kept for tech and gaming pages, and a real-browser checker and finish gate.
1 skill
Open-source AI job search that runs locally in your AI coding CLI: scan job portals, evaluate listings into a structured A-H report with a 1-5 score, tailor your CV, track applications
5 skills · hooks
This page updates itself from your repo. Add the badge to your README, or get it Featured to pin it above the organic results with a followed link to your site.