Pp Carstay logo

Pp Carstay

CommunityPopular
mvanhorn
pp-carstay

Plan Japan designated overnight stops through Carstay: discover Japanese listings, compare facilities, check parking-space constraints, audit evidence and produce booking handoffs. Use for Carstay overnight spots, facilities, Japanese coverage, or carstay-pp-cli.

Overview

Publishermvanhorn
Repositoryprinting-press-library
Skill namepp-carstay
Stars
2.1K
Forks
624
Bundled files
Instructions only
Links
  • Markdown instructions

    A SKILL.md file the model loads on demand, so it only costs tokens when a request actually matches.

  • Works with any LLM

    AI skills are plain Markdown, not provider-specific code, so this works with GPT, Claude, Gemini, Grok, or a local model.

  • Self-contained

    Everything the model needs lives in the instructions — no extra files to sync.

  • Open source

    Published by mvanhorn on GitHub. Read the source before you install it.

Installation

Install the Pp Carstay AI skill in TypingMind to use it with any LLM, or drop it into another agent that reads SKILL.md.

1

Install in TypingMind

TypingMind installs a skill straight from its GitHub folder — it reads SKILL.md, bundles the resource files, and stores the result locally.

  1. Open the app and go to Plugins → Skills.
  2. Choose "Install from GitHub".
  3. Paste the skill folder URL below and confirm.
    https://github.com/mvanhorn/printing-press-library/tree/main/cli-skills/pp-carstay
  4. Enable the skill in any chat where you want it available.
Plugins → Skills → Add skill → From GitHub URL, then paste the folder URL and press Continue.
2

Install in another agent

Any agent that reads the Agent Skills format can use this skill — copy the folder into that agent's skills directory.

Claude Code — .claude/skills
git clone --depth 1 https://github.com/mvanhorn/printing-press-library.git /tmp/printing-press-library
mkdir -p .claude/skills
cp -r /tmp/printing-press-library/cli-skills/pp-carstay .claude/skills/pp-carstay
Restart Claude Code after copying so it picks up the new skill.

Use it in TypingMind

Enable Pp Carstay in any TypingMind chat and the model takes it from there. Its name and description sit in the system prompt, and the moment a request matches, the model loads the full instructions itself — you never invoke it by hand, and it costs no tokens until it is actually used.

The model loads Pp Carstay on its own as soon as a request matches it.

Works with any AI model

AI skills are plain Markdown instructions rather than provider-specific code, so Pp Carstay is not tied to the model it was written for. Install it once in TypingMind and use it with GPT-5, Claude, Gemini, Grok, DeepSeek, Mistral, Llama, or a local model you run yourself — all on your own API keys.

  • Loaded only when it is needed

    The system prompt carries just the name and description. The instructions are fetched on the first matching request, so an idle skill costs nothing.

  • Switch models mid-chat

    Because the skill is instructions rather than code, changing model does not break it — the next model reads the same SKILL.md.

Skill instructions

This is the SKILL.md content the model loads. Read it before installing — a skill is instructions your model will follow.

Carstay overnight planning

Prerequisites for this source build

Carstay is published in the Printing Press library. Build this source checkout to use the amendment fixes before they merge; catalog installation follows the current published release. Use Go 1.26.6 or newer:

bash
cd library/travel/carstay
go build -trimpath -o carstay-pp-cli ./cmd/carstay-pp-cli
./carstay-pp-cli --version

Verify carstay-pp-cli --version in the runtime that will use this skill. Place the built binary on that runtime's $PATH, or invoke ./carstay-pp-cli from its directory. This checkout retains the published version stamp 2026.10.1; the library assigns the next release version after merge.

The installer below provides the current published catalog release; pending amendment changes require the source build above.

Prerequisites: Install the CLI

This skill drives the carstay-pp-cli binary. You must verify the CLI is installed before invoking any command from this skill. If it is missing, install it first:

  1. Install via the Printing Press installer. It defaults binaries to $HOME/.local/bin on macOS/Linux and %LOCALAPPDATA%\Programs\PrintingPress\bin on Windows:
    bash
    npx -y @mvanhorn/printing-press-library install carstay --cli-only
  2. Verify: carstay-pp-cli --version
  3. Ensure the reported install directory is on $PATH for the agent/runtime that will invoke this skill.

If the npx install fails (no Node, offline, etc.), fall back to a direct Go install (requires Go 1.26.6 or newer). This installs into $GOPATH/bin (default $HOME/go/bin), so add that directory to $PATH instead:

bash
go install github.com/mvanhorn/printing-press-library/library/travel/carstay/cmd/carstay-pp-cli@latest

If --version reports "command not found" after install, the runtime cannot see the binary directory on $PATH. Do not proceed with skill commands until verification succeeds.

Search designated overnight spots using the fuller Japanese source, compare facilities and parking-space dimensions, and preserve unknowns. Dated results are provider-filtered candidates; complete quotes and availability require provider confirmation.

When to Use This CLI

Use for designated Carstay overnight stations around a Japan campervan itinerary, Japanese listing coverage, facility/rule evidence, small shortlist comparisons, explicit parking-space constraints and canonical booking handoff. Public reads need no credentials.

Anti-triggers

Use the provider website/human booking flow for reservations, payments, accounts, exact vacancy and complete dated quotes. Use a routing tool for road distance, driving time or tolls. Preserve original Japanese source text; this CLI does not generate translations or certify that any vehicle/pitch is accepted.

Unique Capabilities

These commands compute bounded evidence views from the public Carstay station source.

Overnight itinerary evidence

  • spots compare — Contrast a small shortlist with facility, price and parking-space evidence.

    Use when itinerary decisions require comparable source evidence instead of a raw record.

    bash
    carstay-pp-cli spots compare 632c59b82b614b99a252d1b2 5cff4813839680041631c452 --agent
  • spots fit — Check explicit space dimensions and facility requirements while preserving unknowns.

    Use when itinerary decisions require comparable source evidence instead of a raw record.

    bash
    carstay-pp-cli spots fit 632c59b82b614b99a252d1b2 --length-m 6 --width-m 2.1 --require electricity,restroom --agent
  • spots near — Rank designated overnight spots around an itinerary coordinate by straight-line distance.

    Use when itinerary decisions require comparable source evidence instead of a raw record.

    bash
    carstay-pp-cli spots near --lat 35.5 --lon 138.75 --radius-km 60 --limit 5 --agent
  • spots audit — Expose unresolved price, availability and qualified facility evidence before booking.

    Use when itinerary decisions require comparable source evidence instead of a raw record.

    bash
    carstay-pp-cli spots audit 632c59b82b614b99a252d1b2 --check-in 2026-10-10 --check-out 2026-10-11 --agent

Japanese source coverage

  • spots coverage — See Japanese options and English publication or translation gaps.

    Use when itinerary decisions require comparable source evidence instead of a raw record.

    bash
    carstay-pp-cli spots coverage --prefecture Yamanashi --agent

Optional learning evidence

Manual pattern teaching uses its current resource type, venue, entity kind and examples. Inference and undo use only positive boost teachings from eligible sources; hide and alias_of rows stay explicit local rules and never support a synthesized boost. Undo retains an inferred rule only when its remaining distinct positive bindings still support it. Recall keeps a known conflicting cached resource in the mismatch evidence. Synthesized patterns validate only the entity actually substituted; another entity mentioned by the query cannot validate that target. A pattern tries later query bindings after a cached identity conflict and contributes one valid binding. Recall keeps the best identity evidence for each resource type and ID, then applies its result limit; final returned targets do not also appear as rejected alternatives. Ordinary recall ranks pattern metadata by its final confidence order and stops identity reads once enough exact targets are settled. Direct teachings remain scanned; matching pattern metadata is scanned when pattern evaluation is needed, and --debug-mismatches may validate remaining patterns to collect rejected evidence. Standalone pattern application retains its first ID-verified binding, score ordering and caps. An unavailable cached identity retains identifier-only pattern fallback and proves no provider facts. Only a genuinely missing row uses the warned missing-resource fallback; cancellation and unexpected SQLite payload-read errors fail explicitly.

bash
carstay-pp-cli recall 'Yamanashi overnight stop' --debug-mismatches --json

This optional helper may migrate or record local learning state; native spots provider evidence remains independent. The MCP mirror advertises these local effects. For a failed native MCP shortlist call, inspect isError and its first bounded evidence block before reading separate diagnostics.

Command Reference

Start with carstay-pp-cli spots --help and the relevant leaf help. find supports text, prefecture, language and dates. show keeps original source facts. Use compare to align a 2–5 ID shortlist, fit to test explicit requirements, and audit to inspect unresolved evidence. near ranks by straight-line kilometers, coverage counts locale gaps, and handoff prints the canonical link plus JST date query.

Freshness Contract

spots commands read current public HTTP data and reject --data-source local. A dated result is provider_date_filtered_candidate, with availability=unknown. Its absence inside a scan cap is not unavailability. Starting prices are JPY references/night; option fee units and complete totals remain unknown. Parking dimensions are meters and represent parking-area size, not vehicle acceptance.

Preserve source_present true/false/null and facility notifications. Missing fields are unknown; false plus notification requires confirmation. On-site restroom differs from nearby toilet. Source option price zero with a fee note is not a free-service promise. Nearby numeric distances have an unverified source unit. Japanese is the fuller discovery baseline; English approval and actual translated-name presence are independent.

Recipes

Compare overnight options

bash
carstay-pp-cli spots compare 632c59b82b614b99a252d1b2 5cff4813839680041631c452 --agent

Read a bounded evidence matrix, with starting prices.

Find stops near an itinerary point

bash
carstay-pp-cli spots near --lat 35.5 --lon 138.75 --radius-km 60 --limit 5 --agent

Distance is straight-line kilometers, not driving distance.

Project Japanese coverage

bash
carstay-pp-cli spots coverage --prefecture Yamanashi --agent --select results.japanese_overnight,results.english_approved_overnight

Keep only decision fields in agent output.

Audit a dated candidate

bash
carstay-pp-cli spots audit 632c59b82b614b99a252d1b2 --check-in 2026-10-10 --check-out 2026-10-11 --agent

Check missing or qualified evidence without claiming vacancy.

Auth Setup

Public discovery needs no credentials. Complete bookings in Carstay’s canonical website flow.

Run carstay-pp-cli doctor to verify setup.

Agent Mode

Use --agent; one {meta, results} envelope preserves source/observation/date/scan meaning, and empty collections are []. Use --select results.id,results.name_ja,results.source_url for decision rows; selected agent results retain provenance. --json and --dry-run are also supported.

Bound work independently of output: --limit default 10/max 50; directory --max-scan-records default 500/max 5000; dated --max-scan-pages default 2/max 10; compare/fit/audit at most 5 IDs. --text-limit bounds source text per field (default 2000/max 12000 characters); truncation links back to the original page. Inspect scan_complete, matched_within_scan, source_text_truncated and meta.fetch_failures before making a complete-coverage claim.

Metadata-only agent selection keeps {meta, results} with results=[] and meta.results_omitted_by_select=true. Compare/fit/audit return nonzero when any requested detail read fails; partial rows retain detail_fetch_complete=false, requested/successful counts and fetch_failures.

HTTP 429 is a source error, never zero inventory. Every public client call honors root --timeout; read pacing is capped at 2 requests/second. Source-contract errors remain failures rather than empty results.

MCP companion inputs are scalar slots: compare requires first and second; fit/audit require id; additional IDs use second/third/fourth/fifth where applicable. Show/handoff require id; near requires lat/lon. Do not pass a space-joined variadic ID string.

Cached-resource alias recall requires one canonical shared by the query, teaching and actual cached resource; a conflicting target remains a mismatch. Missing-resource fallback keeps its explicit warning.

Optional learning reads (recall, learnings list/candidates/stats, playbook list) are advertised as local writes in MCP: they keep their current store migration, event or pruning behavior and return current data or an explicit store-unavailable error while another SQLite writer holds WAL mode; after that writer closes, recall sees the committed data. Native read-only MCP mirrors enforce a child context that suppresses implicit journals, flag derivation, refresh/cache writes and receipts. Normal CLI optional journaling remains enabled unless --no-learn is selected; that switch does not make every learning-store open read-only. Carstay station requests are public GETs.

Direct Use

A separate generated reference mirror supports sync --resources directory and offline search --type directory. It can contain activity-only rows and stale reference data; spots handles current overnight evidence. Read README.md for local installation, mirror examples and troubleshooting.

The MCP companion preserves bounded partial CLI evidence in the first content block when a tool fails, keeps isError=true, and separates diagnostics. Inspect both the failed-tool flag and meta.fetch_failures/coverage; partial rows never certify a complete shortlist.

The optional local learning engine requires a unique literal prefix match before treating a cached resource ID as verified. Explicit pattern teaching records its current resource type, venue, entity kind and examples; later inference preserves that manual payload. Teaching undo atomically reconciles affected inferred query/resource/venue families: rules with at least two distinct compatible retained bindings survive with examples from those supporters; stale or incompatible rows do not veto the compatible cohort, unsupported rules are removed, and explicit taught or unrelated patterns remain. Native spots reads use current provider evidence independently of these local learnings.

The generated directory client refuses foreign effective-origin redirects before any inherited configured/per-call headers or cookies are sent. Same-origin normalized default ports, fresh auth signing and the ten-hop limit remain; policy refusals return directly. Existing binary body/MIME and stream timeout rules stay separately scoped. Native spots requests do not use configured credential headers; their source URLs are canonical provider references and do not attest the final redirect origin.

Frequently asked questions

What does the Pp Carstay AI skill do?

Plan Japan designated overnight stops through Carstay: discover Japanese listings, compare facilities, check parking-space constraints, audit evidence and produce booking handoffs. Use for Carstay overnight spots, facilities, Japanese coverage, or carstay-pp-cli.

Why use Pp Carstay on TypingMind?

Because you install it once and use it with any model. Pp Carstay is plain Markdown rather than provider-specific code, so the same skill runs on GPT-5, Claude, Gemini, Grok, or a local model — and you can switch model mid-chat without it breaking. TypingMind runs on your own API keys, so you pay providers directly instead of a per-seat subscription, and your skills and chats stay in your own storage.

How do I install Pp Carstay in TypingMind?

Open Plugins → Skills → Install from GitHub in TypingMind and paste https://github.com/mvanhorn/printing-press-library/tree/main/cli-skills/pp-carstay. TypingMind reads its SKILL.md and installs it as a skill you can enable per chat.

Which AI models can use Pp Carstay?

Any model you connect in TypingMind. AI skills are plain Markdown instructions rather than provider-specific code, so GPT, Claude, Gemini, Grok, and local models can all load this skill when a request matches it.

How many AI models can I use with Pp Carstay?

As many as you like. As long as a model supports skills, you can use Pp Carstay with it — GPT, Claude, Gemini, Grok, DeepSeek, Mistral, Llama and more — all on TypingMind with your own API keys.

Is the Pp Carstay AI skill free?

It is published on GitHub by mvanhorn. Check the repository for licensing terms. You only pay your own AI provider for the tokens you use.

What are AI skills?

An AI skill is a reusable instruction bundle that teaches an AI model how to do one specific task. It follows the open Agent Skills format: a SKILL.md file with a name and description, plus any scripts, templates or reference files the model may need. The model reads the instructions only when your request matches the skill, so an installed skill costs nothing until it is used.

How are AI skills different from plugins or MCP servers?

A plugin or MCP server gives a model new tools to call — code that runs somewhere and returns a result. An AI skill gives the model knowledge and process instead: how to approach a task, which steps to follow, what good output looks like. Skills are plain Markdown, so they need no server, no API key and no runtime, and they work with any model.

View all

Set up your own AI workspace now

Get notified about new features and future giveaways by subscribing to our newsletter 👇