--- name: pp-tabiwa description: "Compare regional tabiwa catalog products with payment units and original restriction evidence. Trigger phrases: `compare tabiwa passes`, `tabiwa points-only ticket`, `tabiwa catalog restrictions`, `use tabiwa`, `run tabiwa-pp-cli`." author: "zjsng" license: "Apache-2.0" argument-hint: " [args] | install cli|mcp" allowed-tools: "Read Bash" metadata: openclaw: requires: bins: - tabiwa-pp-cli install: - kind: go bins: [tabiwa-pp-cli] module: github.com/mvanhorn/printing-press-library/library/travel/tabiwa/cmd/tabiwa-pp-cli --- # tabiwa by WESTER catalog — Printing Press CLI ## Catalog scope and evidence Public catalog reads work without login. The only sent cookie is `regionId`, the website's display preference: `10` せとうち, `20` 北陸, `30` 山陰, `40` 九州. Full detail pages enter Queue-it for ordinary HTTP and are excluded from runtime access. The CLI never imports queue/account cookies, books, pays, retrieves purchased tickets, or redeems a ticket. Use `geography list --region 20` for exact provider prefecture/area IDs; these differ from national prefecture codes. Use `catalog search` for a bounded shortlist, `catalog inspect ID` for the selected catalog summary and `catalog compare ID ID --on YYYY-MM-DD` for price units and requested-date membership. `catalog` and `geography` are bounded aliases of their search/list commands. Live catalog commands support `--data-source auto` or `live`; `catalog saved` supports `auto` or `local`. `price.unit` separates `JPY`, `WESTER_POINT`, and `unknown`. `points_only` is the source boolean or null; a yen display does not establish that points cannot be used. Amounts remain decimal strings. A quote's passenger/unit basis is unknown unless the overview explicitly identifies it, such as a per-vehicle taxi price. The CLI does not rank incomparable quotes, convert points, or calculate pass savings. Requested-date membership is a published catalog filter. `not_listed` does not mean sold out, closed, ineligible or unavailable. Real availability is always unknown. An explicit overview sold-out notice has its own `published_status`; generic limited-stock wording never sets it. Area tags do not prove included routes: J0000900 tags 直島 in its broader area while its overview excludes unlisted 直島. Preserve both facts. `restriction_evidence` contains original Japanese overview lines with restriction/redemption cues, not a complete policy or route description. Complete redemption terms and included routes remain unknown. Read `overview_truncated` and `restriction_evidence_truncated` before relying on the excerpts. Canonical product links lead to the source's final terms; opening them may require its ordinary queue. Source reads have an 8 MiB response/1000-record bound. Search scans at most `--max-records` (default 500, maximum 1000) independently from `--limit` (default 10, maximum 50). Coverage reports source/scanned/returned counts and truncation. A zero result applies only to those filters and scanned records. Inspect requests one ID; compare requests two to five distinct IDs and uses at most two source reads. Every observation retains source URL and UTC retrieval time; requested dates are Japan calendar dates. Generic root `search` uses local FTS over synced geography in auto/local modes; explicit live is unsupported. Use `catalog search` for bounded regional products. Generic MCP search/SQL use geography from the separate framework store. Save-capable catalog MCP tools are local writes because `save=true` persists selected private evidence; provider requests remain GET-only. Geography and saved-evidence tools remain reads. Live commands save only with `--save`; `--no-cache --save` is refused. Selected normalized evidence is stored in `catalog/saved.db` under the resolved cache directory, at most 50 `(region,id)` observations, 24 KiB per record. Only a strictly newer observation instant replaces that product's prior evidence; delayed or equal-clock saves keep the existing observation. SQLite saves use normal transactions. `catalog saved` opens only an existing read-only database and performs no migration or source read; missing storage returns an empty saved list without creating directories. Saved clocks are original source clocks, not current availability. SQL's generic framework store is separate from this selected-evidence store. Queue redirects, HTML challenges, throttles, contract changes, and network failures are explicit errors. Date-comparison failure returns an error instead of a partly fabricated comparison. No automatic source-access workaround or browser runtime is used. A large MCP result may require fewer products or narrower `--select` fields. ## Prerequisites: Install the CLI This skill drives the `tabiwa-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 `$HOME/.local/bin` on macOS/Linux and `%LOCALAPPDATA%\Programs\PrintingPress\bin` on Windows: ```bash npx -y @mvanhorn/printing-press-library install tabiwa --cli-only ``` 2. Verify: `tabiwa-pp-cli --version` 3. Ensure the reported install directory is on `$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 `$PATH` instead: ```bash go install github.com/mvanhorn/printing-press-library/library/travel/tabiwa/cmd/tabiwa-pp-cli@latest ``` If `--version` reports "command not found" after install, the runtime cannot see the binary directory on `$PATH`. Do not proceed with skill commands until verification succeeds. Inspect published catalog summaries and requested-date membership. Full ticket terms, route coverage and stock remain unknown; finish on the canonical product page. ## When to Use This CLI Use for tabiwa regional transport/attraction bundles and coupons, points-only prices, overview restriction evidence and saved catalog shortlists. ## Anti-triggers Do not use this CLI for: - Do not use for bookings, payments, purchased tickets, rewards balances or redemption actions. - Do not infer real availability, full redemption policies, included routes or savings from catalog labels. ## Unique Capabilities These capabilities aren't available in any other tool for this API. ### Catalog decisions - **`catalog search`** — Discover regional products with a dated catalog filter and explicit scanned coverage. _Discover regional products with a dated catalog filter and explicit scanned coverage._ ```bash tabiwa-pp-cli catalog search --region 20 --category transportation --on 2026-10-28 --limit 3 --agent ``` - **`catalog compare`** — Compare points-only and yen-quoted products without inventing currency conversion. _Compare points-only and yen-quoted products without inventing currency conversion._ ```bash tabiwa-pp-cli catalog compare J0001900 J0000900 --region 10 --agent ``` - **`catalog inspect`** — Read original Japanese restriction and redemption hints while complete terms remain unknown. _Read original Japanese restriction and redemption hints while complete terms remain unknown._ ```bash tabiwa-pp-cli catalog inspect J0000900 --region 10 --agent ``` - **`catalog compare`** — Show whether selected products appear in a requested-date catalog without claiming ticket stock. _Show whether selected products appear in a requested-date catalog without claiming ticket stock._ ```bash tabiwa-pp-cli catalog compare J0001900 J0000900 --region 10 --on 2026-10-28 --agent ``` - **`catalog saved`** — Read saved selected products offline with their original observation times. _Read saved selected products offline with their original observation times._ ```bash tabiwa-pp-cli catalog saved --region 10 --limit 3 --agent ``` ## Command Reference **catalog** — Inspect published ticket catalog summaries and their limits - `tabiwa-pp-cli catalog` — Discover catalog summaries with regional and date filters; not live stock **geography** — Resolve provider region-specific prefecture and area identifiers - `tabiwa-pp-cli geography` — List source prefecture and area IDs for a selected regional catalog ### Finding the right command When you know what you want to do but not which command does it, ask the CLI directly: ```bash tabiwa-pp-cli which "" ``` `which` resolves a natural-language capability query to the best matching command from this CLI's curated feature index. Exit code `0` means at least one match; exit code `2` means no confident match — fall back to `--help` or use a narrower query. `--json` (and other machine formats) keep that exit-2 contract and write `{"matches":[]}` on stdout so agents can inspect the envelope without treating a miss as success. ## Recipes ### Payment units ```bash tabiwa-pp-cli catalog compare J0001900 J0000900 --region 10 --agent ``` Compare points-only and yen-quoted products without inventing currency conversion. ### Source restrictions ```bash tabiwa-pp-cli catalog inspect J0000900 --region 10 --agent ``` Read original Japanese restriction and redemption hints while complete terms remain unknown. ### Compact discovery ```bash tabiwa-pp-cli catalog search --region 20 --category transportation --limit 3 --agent --select products.id,products.name,products.price,coverage ``` Keep identity, unit-qualified quotes and scan coverage. ## Auth Setup Public catalog reads need no login. The only sent cookie is the documented regionId display preference. Queue-it full detail pages are excluded from runtime access. Run `tabiwa-pp-cli doctor` to verify setup. ## Agent Mode Add `--agent` to any command. Expands to: `--json --compact --no-input --no-color`. Global format flags share one contract on promoted, novel, sync, and `--deliver` paths: - `--json` — one JSON document on stdout (sync progress events go to stderr) - `--compact` — keep identity/status/timestamp fields; does not change the document vs stream shape - `--csv` / `--plain` — tabular rows (collection envelopes unwrap to the row array) - `--quiet` — one identity value per row, no envelope - **Pipeable** — JSON on stdout, errors on stderr - **Filterable** — `--select` keeps a subset of fields. Dotted paths descend into nested structures; arrays traverse element-wise. Critical for keeping context small on verbose APIs: ```bash tabiwa-pp-cli catalog --agent --select id,name,price ``` - **Previewable** — `--dry-run` shows the request without sending - **Offline-friendly** — sync/search commands can use the local SQLite store when available - **Non-interactive** — never prompts, every input is a flag - **Read-only** — do not use this CLI for create, update, delete, publish, comment, upvote, invite, order, send, or other mutating requests ### Response envelope Commands that read from the local store or the API wrap output in a provenance envelope: ```json { "meta": {"source": "live" | "local", "synced_at": "...", "reason": "..."}, "results": } ``` Parse `.results` for data and `.meta.source` to know whether it's live or local. A human-readable `N results (live)` summary is printed to stderr only when stdout is a terminal AND no machine-format flag (`--json`, `--csv`, `--compact`, `--quiet`, `--plain`, `--select`) is set — piped/agent consumers and explicit-format runs get pure JSON on stdout. ## Paths and state Agents should treat the CLI's path resolver as part of the runtime contract: - Use `--home ` for one invocation, or set `TABIWA_HOME=` to relocate all four path kinds under one root. - Use per-kind env vars only when a specific kind must diverge: `TABIWA_CONFIG_DIR`, `TABIWA_DATA_DIR`, `TABIWA_STATE_DIR`, `TABIWA_CACHE_DIR`. - Resolution order is per-kind env var, `--home`, `TABIWA_HOME`, XDG (`XDG_CONFIG_HOME`, `XDG_DATA_HOME`, `XDG_STATE_HOME`, `XDG_CACHE_HOME`), then platform defaults. - `config` contains `config.json` and profiles. `data` contains the generic framework `data.db`. `state` contains local runtime and learning files. `cache` contains regenerable HTTP/cache files and selected observations in `catalog/saved.db`. - Run `tabiwa-pp-cli doctor --fail-on warn` to surface path warnings. `agent-context` exposes a schema v4 `paths` block for agents that need the resolved dirs. - For MCP, pass relocation through the MCP host config. The MCP binary does not inherit CLI flags: ```json { "mcpServers": { "tabiwa": { "command": "tabiwa-pp-mcp", "env": { "TABIWA_HOME": "/srv/tabiwa" } } } } ``` Fleet precedence: an inherited per-kind env var overrides an explicit `--home` for that kind. Use `TABIWA_HOME` or per-kind vars as durable fleet levers, and use `--home` only for a single invocation. Relocation is not reversible by unsetting env vars; move files manually before clearing `TABIWA_HOME`, or `doctor` will not find credentials left under the former root. ## Automatic learning This CLI ships a self-capturing learning loop. The CLI does its own bookkeeping: every invocation is journaled locally, a failed flag followed by a corrected retry auto-derives a `flag_alias` candidate, and a `teach` on a query family without a playbook auto-synthesizes a `playbook_candidate` from the session's journal. Your job is judgment only: `recall` first, act on surfaced candidates, `teach` the final answer, `playbook amend` when you observe a correction. You never record failures by hand. ### Step 1: `recall` before any discovery Before list/search/drill commands on a new user question, pass the question as an argv or MCP tool argument to `recall --agent`. Do not interpolate user-controlled text into a shell command line. Quoted `recall ""` breaks on an apostrophe, which is ordinary English. A quoted heredoc breaks when a body line equals the delimiter, and that delimiter is published in these docs. Write the question with a non-shell file-writing tool, then read it back as data: ```bash # Write the question verbatim with your file-writing tool (no shell involved). # Command substitution on a file only ever yields data — the shell never # parses the file's bytes as syntax. QUERY=$(cat /path/to/question.txt) tabiwa-pp-cli recall "$QUERY" --agent ``` Prefer MCP: pass the question as the tool's query argument. `"$QUERY"` after a file read is argv-safe; putting the question itself in the command text is not. The response envelope: ```json { "query": "...", "normalized": "", "query_entities": ["..."], "found": true | false, "match_score": 0.0, "results": [ { "resource_id": "...", "resource_type": "...", "venue": "...", "confidence": 2, "entity_match": "exact|partial|unknown", "source": "taught|preseed|pattern", "warnings": ["..."] } ], "mismatches": [ /* only when --debug-mismatches */ ], "warnings": [ /* top-level */ ], "candidates": [ { "id": 12, "class": "flag_alias | playbook_candidate", "summary": "...", "sightings": 3, "last_seen": "...", "rationale": "...", "next_action": ["", "tabiwa-pp-cli learnings confirm 12"] } ], "playbook": { "query_family": "...", "playbook": { "steps": [ { "cmd": "", "purpose": "..." } ], "entity_slots": ["$ENTITY"], "expected_tool_calls": 3 }, "slots_resolved": { "$ENTITY": { "token": "", "canonical": "" } }, "notes": "" }, "notes": "" } ``` Empty-store short-circuit: if the store has no learnings, playbooks, or candidates yet (recall finds nothing and `learnings list` and `learnings candidates` are both empty), skip recall for the rest of this session instead of taxing every query; resume recall-first once something has been taught. ### Step 2: decision tree Read `candidates`, `playbook`, `notes`, `results[0]`, and warnings in that order: ``` if Candidates present (warnings include "candidates_present"): -> candidates are try-then-confirm, never facts. Follow each candidate's two-step next_action verbatim: run the trial command first, then run `learnings confirm ` only after the trial verified the behavior. Reject a wrong candidate with `learnings reject `. -> NEVER re-teach something recall surfaced as a candidate; confirm or reject that candidate instead of teaching a duplicate. -> candidates ride alongside playbooks and resource hits, not instead of them; continue with the branches below after acting on them. if Playbook present: -> READ Playbook.notes verbatim FIRST (workarounds + gotchas the CLI surface doesn't expose) -> replay Playbook.steps in order, substituting Playbook.slots_resolved entries for the entity slot tokens. If a step's slot is unresolved, fall back to discovery for that step only. -> the Playbook's expected_tool_calls is a budget; if you find yourself running materially more, record the divergence via `tabiwa-pp-cli playbook amend` at end-of-session. elif Notes present (no Playbook): -> read Notes verbatim before any discovery step; they carry known gotchas for this query family even when no structured choreography exists yet. elif Found AND Results[0].EntityMatch == "exact" AND Results[0].Confidence >= 2: -> skip discovery; fetch live data for Results[*].ResourceID in parallel elif Found AND Results[0].EntityMatch == "partial": -> candidate hint, NOT a hit; read the resource title to validate before trusting elif (any row in Mismatches[] when --debug-mismatches was passed): -> treat as cold start; the stored learning is for a different entity (different canonical resolved from query_entities) else: // Found == false, no playbook, no notes -> cold start; run discovery normally; teach the answer afterward (Step 4). If the family has no playbook yet, that teach auto-synthesizes a playbook candidate from this session's journal - you do not need to record one by hand. ``` Playbook and Notes are orthogonal to the per-resource path. A recall response can carry both a Playbook AND a `Results[]` hit - use both: the Playbook tells you which choreography to run; the resource hits short-circuit specific steps. Default to skipping `mismatches`; pass `--debug-mismatches` only when investigating cold-start surprises. Candidate judgment details: `learnings confirm ` prints the candidate's full payload before materializing it - check that the printed payload matches the behavior you verified. `learnings reject ` tombstones the derivation signature so the same candidate does not resurface. The envelope carries only the few candidates worth acting on now; `tabiwa-pp-cli learnings candidates` lists the full open set. Graceful degradation: if `learnings confirm` is an unknown command, you are driving an older binary - ignore the candidates guidance and follow the rest of the protocol. ### Step 3: always read `warnings` - `low_confidence`: row exists at `confidence<2`. Treat as a hint, not a skip-discovery hit. - `resource_not_in_store`: the local store doesn't have the resource the learning points at. The match validator couldn't classify entities — direct-fetch an