--- name: pp-irail description: "Every Belgian rail lookup the existing tools offer, plus transfer-risk, delay history and accessibility data that none of them have. Trigger phrases: `next train from Brussels-Central`, `when is my train to Ghent`, `is my train delayed`, `are there any rail disruptions in Belgium`, `plan a train journey from Leuven to Antwerp`, `will I make my transfer`, `use irail`, `run irail`." author: "Olivier" license: "Apache-2.0" argument-hint: " [args] | install cli|mcp" allowed-tools: "Read Bash" metadata: openclaw: requires: bins: - irail-pp-cli install: - kind: go bins: [irail-pp-cli] module: github.com/mvanhorn/printing-press-library/library/travel/irail/cmd/irail-pp-cli --- # iRail — Printing Press CLI ## Prerequisites: Install the CLI This skill drives the `irail-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 `$HOME/.local/bin` on macOS/Linux and `%LOCALAPPDATA%\Programs\PrintingPress\bin` on Windows: ```bash npx -y @mvanhorn/printing-press-library install irail --cli-only ``` 2. Verify: `irail-pp-cli --version` 3. Ensure the reported install directory is on `$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 1.26.5 or newer). This installs into `$GOPATH/bin` (default `$HOME/go/bin`), so add that directory to `$PATH` instead: ```bash go install github.com/mvanhorn/printing-press-library/library/travel/irail/cmd/irail-pp-cli@latest ``` If `--version` reports "command not found" after install, the runtime cannot see the binary directory on `$PATH`. Do not proceed with skill commands until verification succeeds. iRail exposes live NMBS/SNCB departures, journey planning and disruptions for free with no API key. This CLI adds what every other client throws away: it records observations locally so punctuality can answer whether a train is chronically late, joins the open stations dataset so transfer-risk knows the real minimum transfer time at each station, and surfaces station accessibility data the API never returns. The analysis commands emit typed JSON - delays as numbers, cancellations as booleans - while the raw endpoint commands pass iRail's payload through unchanged. ## When to Use This CLI Use this CLI for anything involving Belgian passenger rail: when the next train leaves, how to get from one station to another, whether a journey is disrupted, and how reliable a route has been over time. It is the right tool when you need typed, scriptable output rather than a web page, and the only tool that can answer historical punctuality or transfer-risk questions. It also serves accessibility questions about Belgian stations that the live API cannot answer. ## Anti-triggers Do not use this CLI for: - Do not use this CLI to buy tickets, reserve seats or check fares; iRail is a read-only timetable API with no booking surface. - Do not use it for rail networks outside Belgium; coverage is NMBS/SNCB plus a small number of cross-border stations. - Do not use it for live vehicle GPS positions; the API reports scheduled and delayed times, not continuous tracking. - Do not use it to query personal NMBS account data such as subscriptions or order history; that lives behind belgiantrain.be, which this CLI does not touch. ## Unique Capabilities These capabilities aren't available in any other tool for this API. ### Live data joined with open datasets - **`transfer-risk`** — Tells you whether the transfers in a journey still hold once today's delays are applied. _Reach for this instead of a plain route lookup whenever a journey has a transfer and delays are already in play._ ```bash irail-pp-cli transfer-risk --from Oostende --to Hasselt --agent ``` - **`disruptions route`** — Filters the national disruption feed down to the stations your journey actually passes through. _Use this when the national list is too noisy to answer whether one specific trip is affected._ ```bash irail-pp-cli disruptions route --from Ghent-Sint-Pieters --to Brussels-Central --agent ``` - **`stations facilities`** — Reports step-free access, elevators, ramps, lockers, bike parking and ticket-desk hours for a station. _Use this for accessibility and amenity questions that the live rail API simply cannot answer._ ```bash irail-pp-cli stations facilities --station Ghent-Sint-Pieters --agent ``` ### Local state that compounds - **`punctuality`** — Shows how reliable a train or route has actually been, from delay observations recorded on your machine. _Use this for questions about the past such as chronic lateness; it never calls the API._ ```bash irail-pp-cli punctuality --from Ghent-Sint-Pieters --to Brussels-Central --board-type route --agent ``` - **`observe`** — Records what the board says right now into local SQLite, building the history other commands read. _Run this on a schedule; it is what makes punctuality and changes able to answer anything._ ```bash irail-pp-cli observe --station Brussels-Central ``` - **`changes`** — Reports new delays, cancellations and platform changes since the last time you looked. _Use this for deltas during a commute rather than re-reading a whole board._ ```bash irail-pp-cli changes --station Brussels-Central --agent ``` ### Time reasoning done properly - **`leave-by`** — Answers the last train you can take and still arrive before a deadline, accounting for current delays. _Use this when the arrival deadline is fixed and the departure time is the unknown._ ```bash irail-pp-cli leave-by --from Leuven --to Brussels-Central --arrive-by 09:00 --agent ``` ## Command Reference **board** — Live departure and arrival boards for a station - `irail-pp-cli board` — Live departures (or arrivals) at a station, with delay, platform and occupancy **disruptions** — Network-wide disruptions and planned engineering works - `irail-pp-cli disruptions` — Current disruptions and planned works across the whole network **logs** — Recent iRail API request log entries - `irail-pp-cli logs` — Last request log entries. Note: upstream currently returns an empty list; bulk archives live at gtfs.irail.be/logs/ **route** — Journey planning between two stations, with transfers and live delay - `irail-pp-cli route` — Plan a journey between two stations, including transfers and per-leg delay **stations** — Belgian and cross-border stations served by NMBS/SNCB - `irail-pp-cli stations` — List every station iRail knows about (716 incl. cross-border) **train** — Individual trains: full stop trace and physical composition - `irail-pp-cli train composition` — Physical make-up of a train: segments, units, and per-carriage facilities - `irail-pp-cli train get` — Every stop of one train with live delay. Note: iRail ignores the date parameter ### Finding the right command When you know what you want to do but not which command does it, ask the CLI directly: ```bash irail-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. ## Recipes ### Next departures, narrowed for an agent ```bash irail-pp-cli board --station Brussels-Central --agent --select vehicle,delay,platform,canceled ``` A full board is roughly 34 KB of JSON; selecting four fields keeps the response small enough to reason over without burning context. ### Will my transfer survive today's delays ```bash irail-pp-cli transfer-risk --from Oostende --to Hasselt --agent ``` Joins live per-leg delay against each station's official minimum transfer time and flags transfers that no longer hold. ### Only the disruptions that affect my commute ```bash irail-pp-cli disruptions route --from Ghent-Sint-Pieters --to Brussels-Central --agent ``` Filters the national feed, which was 32 entries on a normal day, down to the stations this journey passes through. ### Last train that still gets me there by nine ```bash irail-pp-cli leave-by --from Leuven --to Brussels-Central --arrive-by 09:00 --agent ``` Plans backwards from the deadline and applies current delays plus a safety margin. ### Is this station step-free ```bash irail-pp-cli stations facilities --station Ghent-Sint-Pieters --agent ``` Reads the open facilities dataset for elevators, ramps and wheelchair access, none of which the rail API returns. ## Auth Setup No credentials are required: the iRail API is open and unauthenticated, so every command works immediately after install. Two operational limits matter instead of a key. iRail allows 3 requests per second per IP with 5 burst, returning HTTP 429 beyond that, so this CLI ships an adaptive rate limiter. iRail also blocks source IPs that send no User-Agent without prior warning, so every request carries an identifying User-Agent automatically. Run `irail-pp-cli doctor` to verify setup. ## Agent Mode Add `--agent` to any command. Expands to: `--json --compact --no-input --no-color --yes`. - **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 irail-pp-cli board --station Ghent-Sint-Pieters --agent --select id,name,status ``` - **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 `IRAIL_HOME=` to relocate all four path kinds under one root. - Use per-kind env vars only when a specific kind must diverge: `IRAIL_CONFIG_DIR`, `IRAIL_DATA_DIR`, `IRAIL_STATE_DIR`, `IRAIL_CACHE_DIR`. - Resolution order is per-kind env var, `--home`, `IRAIL_HOME`, XDG (`XDG_CONFIG_HOME`, `XDG_DATA_HOME`, `XDG_STATE_HOME`, `XDG_CACHE_HOME`), then platform defaults. - `config` contains settings like `config.toml` and profiles. `data` contains `credentials.toml`, `data.db`, cookies, and auth sidecars. `state` contains persisted queries, jobs, and `teach.log`. `cache` contains regenerable HTTP/cache files. - Stored secrets live in `credentials.toml` under the data dir. Existing legacy `config.toml` secrets are read for compatibility and leave `config.toml` on the first auth write. - Run `irail-pp-cli doctor --fail-on warn` to surface path and credential-location 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": { "irail": { "command": "irail-pp-mcp", "env": { "IRAIL_HOME": "/srv/irail" } } } } ``` Fleet precedence: an inherited per-kind env var overrides an explicit `--home` for that kind. Use `IRAIL_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 `IRAIL_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, run: ```bash irail-pp-cli recall "" --agent ``` 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": ["", "irail-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 `irail-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; `irail-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 and re-evaluate. - `cross_alias_match` (per-result): the row was taught under a different alias and matched the live query's canonical via `entity_lookups` (e.g., a "USA" teach satisfying a "United States" recall). Trust the resource_id. - `similar_shape_different_entity:` (top-level): a structurally matching row exists but its canonical entity differs from the live query's. Treated as cold start; the warning carries the conflicting canonical as a hint, but the row is NOT promoted into Results. - `ambiguous_alias` (top-level): a single query entity resolved to multiple canonicals (e.g., "Cards" → Arizona Cardinals + St. Louis Cardinals). Surface the ambiguity from context before committing to a resource. - `candidates_present` (top-level): the envelope carries a `candidates` section. Handle it via the candidates branch in Step 2 before anything else. - `lookup_refresh_available` (top-level): an entity in the query has no lookup row yet, but synced data could provide one. Run `irail-pp-cli sync` to refresh entity lookups. - Top-level `no_learnings_for_query_family`: the table had no rows above the Jaccard floor. Pure cold start. ### Step 4: `teach &` after finalizing your response - always Teaching is unconditional. After resolving a query the store could not answer, background-teach the final resource mapping - no call-count threshold, no judging whether it was "worth" learning. The teach is the anchor of the loop: it triggers playbook synthesis for a family without a playbook, and same-referent phrasings fold into one family so near-duplicate teache