Pp Concur

SAP Concur — Printing Press CLI

Prerequisites: Install the CLI

This skill drives the concur-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:

   bashCopy code
   npx -y @mvanhorn/printing-press-library install concur --cli-only
   

2. Verify: concur-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:

bashCopy code
go install github.com/mvanhorn/printing-press-library/library/productivity/concur/cmd/concur-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.

SAP Concur's official API requires enterprise partner credentials most individual users can never get. This CLI defaults to your logged-in browser session instead, so filing expense reports and checking travel works the same day you install it. Local SQLite sync turns your report history into something you can search, join, and validate offline.

When to Use This CLI

Use this CLI for filing and checking your own SAP Concur expense reports and travel -- creating reports, pulling in available card charges, validating and submitting, and checking trip/itinerary status. It is the right choice whenever the task is something an individual employee would otherwise do by logging into concursolutions.com.

Anti-triggers

Do not use this CLI for:

• Do not use this CLI to register a new SAP Concur partner application or manage OAuth2 App Center listings -- that is an admin/partner workflow, not an end-user one.
• Do not use this CLI to actually book new travel (flights/hotels/cars). flights search and hotels search return real, live fares and rates from your actual corporate-negotiated policy for research and comparison, but completing a booking is a complex multi-step web/agent experience this CLI does not implement -- finish the booking in Concur's web app.
• Do not use this CLI for company-wide financial reporting or accounting-system integration -- use the documented v3/v4 partner REST API directly for that.

Unique Capabilities

These capabilities aren't available in any other tool for this API.

Conditional browser fallback for report/expense creation, expense update, and report submission

• reports create — Automatically and transparently retries report creation via automated browser when the Concur v4 API rejects pure HTTP requests with a policyId is required error. This fallback is completely conditional and only triggers for tenants requiring explicit policy assignment. It never guesses a Concur region: the UI host is derived only from a base URL that's actually concursolutions.com, or from an explicit CONCUR_UI_BASE_URL override — anything else is a clear error, not a silent default. If the browser click already succeeded before a later step fails, the error says so explicitly (with the report ID when known) instead of looking like a safely-retryable failure; do not blindly retry in that case.
• expenses create — Same conditional fallback pattern, triggered by a different confirmed-live defect: once a request body passes every client-side validation check, the API 404s instead of persisting (deliberately-invalid bodies correctly get a 400 instead, ruling out a body-shape bug). Triggers for a --stdin body exactly as reliably as the flag-driven path -- the fields it needs are read from the constructed request body, not the command's flag variables. Since the browser never surfaces a usable expense ID, success is confirmed by diffing the report's expense list (GET .../reports/{id}/expenses, confirmed unaffected by the defect) before and after the form's Save click, correlated against the submitted amount and expense type so a report modified concurrently through a shared manager/processor/proxy context can't have another actor's new expense misattributed as this call's own. Fills --vendor/--business-purpose directly in the form when set; if either can't be found or filled, the fallback aborts BEFORE the irreversible Save click rather than saving without a value the caller explicitly asked for. Transaction Date is filled via Concur's calendar picker (day buttons carry a fully descriptive accessible name; direct text entry does not reliably replace an existing value), so any --date is honored, not just today. Payment Type's only verified-live value remains Concur's own default (Cash) -- a --payment-type this can't honor is still rejected before the browser even opens.
• expenses update — Same 404 defect, on the sibling PATCH path (.../expenses/{id}). Fills --purpose/--amount (the latter added specifically to support this fallback) in the expense's direct edit view, then re-fetches the expense and verifies its fields actually match what was requested before reporting success -- a save whose click fired but didn't actually persist is reported as a failure, not trusted just because the follow-up GET itself succeeded. Business Purpose is required on Concur's form, so --purpose "" (a clear request) is rejected up front. --comment/--custom-fields aren't yet verified live through this fallback and are rejected the same way.
• reports submit — A different-shaped (405, not 404) but same-root-cause defect: POST/PUT/DELETE all return 405 and PATCH 404s with the same "No static resource" signature against .../reports/{id}/submit -- no verb reaches the action server-side. Clicks the report page's "Submit Report" button, then a SECOND, separate "Submit Report" button inside the confirmation dialog that opens (both share an identical accessible name and role while the dialog is open -- disambiguated by excluding the already-clicked ref rather than risking a non-deterministic re-click of the wrong one). Confirms isSubmitted/approvalStatus on the re-fetched report before reporting success.

Local state that compounds

• expenses scan-duplicates — Find potential double-entered charges across all of your synced expenses.

  Run this before submitting a batch of reports if you suspect a corporate-card charge and a manually-entered cash expense might be the same transaction.

  bashCopy code
  concur-pp-cli expenses scan-duplicates --agent
  

Live travel shopping (search only, never books)

• flights search — Real flight availability and fares from a live shopping session against your actual corporate-negotiated rates and travel policy -- not a public fare aggregator. One-way by default; --return includes both legs in the search but only renders the outbound leg (see --help for the known gap).

  Use this to compare real options before requesting travel, with policy-compliance flags already applied per fare.

  bashCopy code
  concur-pp-cli flights search --from LAX --to "New York" --depart 2026-10-12 --yes --agent
  

• hotels search — Real hotel availability and rates via a live, policy-scoped search -- the same inventory and pricing Concur's own Hotel Search page shows. Drives a real browser (see HTTP Transport and Auth Setup below) rather than a direct API call, because the hotel shopping-session mutation is blocked from scripted replay by the tenant's bot-mitigation.

  A one-time dedicated-browser setup (see Auth Setup) avoids a separate login every time this command's session expires.

  bashCopy code
  concur-pp-cli hotels search --to "New York" --check-in 2026-10-12 --check-out 2026-10-18 --yes --agent
  

HTTP Transport

This CLI uses Chrome-compatible HTTP transport for browser-facing endpoints. It does not require a resident browser process for normal API calls.

Exception: hotels search. Confirmed live that Concur's hotel shopping-session mutation is blocked from scripted HTTP replay by the tenant's bot-mitigation (byte-for-byte replay of a request that had just succeeded natively in the browser still failed). So this one command drives a real browser via agent-browser instead (npm install -g agent-browser && agent-browser install) -- flights search and every other command remain pure HTTP; only the hotel-shopping mutation needs a real browser.

Command Reference

account — Current user profile, policies, and delegate context

• concur-pp-cli account travel <user_id> — Get the current user's travel profile and loyalty programs
• concur-pp-cli account whoami — Get the current user's profile, addresses, and travel IDs

attendees — Attendee catalog and per-expense attendee associations

• concur-pp-cli attendees add — Add attendees to an expense (merge-preserves existing associations; Concur's underlying association call is a replace
• concur-pp-cli attendees list — Get attendees currently associated with an expense

delegates — Delegate (act-on-behalf-of) relationships

• concur-pp-cli delegates — List users the current session user delegates for, with permission flags

expense_types — Expense type catalog and per-type dynamic form fields

• concur-pp-cli expense-types list — List usable expense types for the current user's policy

expenses — Expense line items within a report

• concur-pp-cli expenses create — Create an expense inside a report (core v3-equivalent fields: type, date, amount, currency, payment type)
• concur-pp-cli expenses get — Get a single expense with its filled/empty field manifest
• concur-pp-cli expenses list — List every expense line item on a report (the only live way to enumerate an existing report's expenses -- reports get returns only the header)
• concur-pp-cli expenses update — Fill or change writable fields on an expense (core + custom/list fields)

flights — Search flight locations, travel policy preferences, and real flight availability (creates a live shopping session -- searches only, never books)

• concur-pp-cli flights locations <query> — Resolve an airport, city, or metro name to Concur's travel location IDs; metro queries (e.g. "New York") resolve to one search endpoint covering all constituent airports
• concur-pp-cli flights preferences — Show your travel policy's flight search defaults
• concur-pp-cli flights search --from <origin> --to <dest> --depart <date> [--return <date>] — Search real flight availability and fares

hotels — Search real hotel availability and rates (drives a real browser search -- searches only, never books)

• concur-pp-cli hotels search --to <destination> --check-in <date> --check-out <date> — Search real hotel availability and rates; requires agent-browser installed (see HTTP Transport and Auth Setup below)

lists — Valid values for list-type expense form fields

• concur-pp-cli lists --list-id <id> — Get valid values for a list-type form field by list ID

locations — Location catalog for filling expense/attendee location fields

• concur-pp-cli locations <query> — Search the location catalog by city or venue name

payment_types — Payment type catalog (Cash, Company Card, etc.)

• concur-pp-cli payment-types — List payment types available to the current user

receipts — Receipt image/PDF attachment

• concur-pp-cli receipts <expense_id> --file <path> — Attach a receipt image or PDF to an expense

reports — Expense report headers and lifecycle

• concur-pp-cli reports create — Create a new expense report header (transparently falls back to browser-driven creation on tenants requiring policy selection)
• concur-pp-cli reports get — Get a report's header
• concur-pp-cli reports list — List the current user's expense reports
• concur-pp-cli reports submit — Submit a report for approval
• concur-pp-cli reports update — Update a report's name or business purpose

requests — Travel requests / pre-trip authorization (UNVERIFIED paths -- see spec header notes)

• concur-pp-cli requests get — Get a travel request's detail and workflow status
• concur-pp-cli requests list — List the current user's travel requests

travel_allowance — Per-diem / travel allowance calculations (UNVERIFIED path -- see spec header notes)

• concur-pp-cli travel-allowance <trip_id> — Get travel allowance (per-diem) calculation results for a trip

trips — Booked trips and itineraries (UNVERIFIED paths -- see spec header notes)

• concur-pp-cli trips get — Get a trip's itinerary detail
• concur-pp-cli trips list — List the current user's upcoming and past trips

Finding the right command

When you know what you want to do but not which command does it, ask the CLI directly:

bashCopy code
concur-pp-cli which "<capability in your own words>"

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

Check for duplicate charges

bashCopy code
concur-pp-cli expenses scan-duplicates --agent

See what's actually on an existing report

bashCopy code
concur-pp-cli expenses list --user-id <uid> --report-id <report_id> --agent

reports get returns only the report header -- this is the only live way to see the report's actual line items: each expense's ID, type, vendor, business purpose, amount, and receiptImageId (check this is non-null to confirm a receipt attachment actually persisted server-side rather than trusting a mutating call's own reported success).

Compare real flight and hotel options before requesting travel

bashCopy code
concur-pp-cli flights search --from LAX --to "New York" --depart 2026-10-12 --yes --agent
concur-pp-cli hotels search --to "New York" --check-in 2026-10-12 --check-out 2026-10-18 --yes --agent

Both create a live shopping session against your real tenant -- searches only, never books. flights search is a direct API call; hotels search drives a real browser (see HTTP Transport and Auth Setup) and is markedly slower.

Scan the local SQLite cache for likely double-entered transactions across all your reports.

File a manual expense line item (e.g. a recurring personal-reimbursement stipend)

bashCopy code
concur-pp-cli expenses create \
  --report-id <report-id> --user-id <user-id> \
  --type 01000 --date 2026-09-15 --amount 50 \
  --payment-type CASH --vendor "F45 Training Culver City" --business-purpose "gym" \
  --agent

--type/--payment-type take the expenseTypeId/paymentTypeId codes from expense-types list/payment-types, not display names. --vendor and --business-purpose are distinct Concur form fields (confirmed live 2026-09-15) -- Vendor Description is who you paid, Business Purpose is why; don't put a business-purpose-shaped value like "gym" into --vendor. Setting --business-purpose at creation time also means never needing expenses apply-rules' PATCH-based fill, which hits this same command's confirmed-live 404 defect just like creation itself used to. --currency set to anything other than USD is rejected -- no working currency-override field is confirmed live for this endpoint, and silently creating an expense in the report/policy default currency instead of what was requested is a correctness bug, not an acceptable fallback. If this command hits its confirmed-live HTTP 404 defect, it now falls back to browser automation automatically instead of just failing -- see "Conditional browser fallback for report/expense creation, expense update, and report submission" above; this triggers for a --stdin body just as reliably as the flag-driven path. Any --date is honored via Concur's calendar picker, but a non-Cash --payment-type this can't reliably honor is still rejected (rather than silently substituted) up front.

Auth Setup

Concur's documented OAuth2 partner API is gated behind a Partner Enablement Manager relationship -- there is no self-serve signup, and this CLI does not implement that OAuth2 flow at all. Instead, this CLI authenticates via cookie/browser-session auth: run 'auth login --chrome', log into your company's Concur portal like you normally would (including SSO/MFA), and the CLI captures the resulting session. If a command fails with 401/403 and your company IT has partner OAuth2 credentials, that workflow requires calling the documented v3/v4 REST API directly (developer.concur.com) outside this CLI -- it is not something 'auth login' or any other command here can switch to.

Run concur-pp-cli doctor to verify setup.

hotels search has a second, separate login by default

hotels search drives its own agent-browser-controlled Chrome instance (see HTTP Transport above), which does not share cookies with auth login --chrome's source browser or credential store. Confirmed live that bridging them by copying cookies does not work -- Concur's bot-mitigation appears to bind the session to the browser/device that created it, not just the cookie value, so a copied JWT gets cleared by the server on the next navigation even when every cookie (including the Akamai bot-sensor ones) is copied alongside it. The first time (or whenever that session expires), hotels search opens its own Chrome window and asks you to log in there directly -- that login persists across later invocations until it expires again, so this is an occasional cost, not a per-search one.

Optional one-time setup to avoid that second login entirely: run a dedicated Chrome profile with remote debugging enabled and log into Concur there once. Use a real named profile (Chrome menu -> "Add Person", or chrome://settings -> Add profile) rather than a throwaway --user-data-dir, so auth login --chrome --profile "<name>" can read its cookies too:

bashCopy code
"/Applications/Google Chrome.app/Contents/MacOS/Google Chrome" \
  --remote-debugging-port=9222 --profile-directory="<profile dir name>"

hotels search auto-detects that session (tries CDP ports 9222, 9333, 9229 in order, or set CONCUR_CDP_PORT for a custom port) and attaches to it -- rather than copying its credentials -- before falling back to its own isolated login. Attaching, not copying, is what makes this work: it is the same live browser connection, so there is no separate device fingerprint for Concur's bot-mitigation to reject. Keep that Chrome window running whenever you plan to use hotels search; if you see "the dedicated Concur browser ... is no longer logged in", log in there again.

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:

  bashCopy code
  concur-pp-cli payment-types --agent --select paymentTypeId,paymentTypeName,description
  

• 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

• Explicit retries — use --idempotent only when an already-existing create should count as success

Response envelope

Commands that read from the local store or the API wrap output in a provenance envelope:

jsonCopy code
{
  "meta": {"source": "live" | "local", "synced_at": "...", "reason": "..."},
  "results": <data>
}

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 <dir> for one invocation, or set CONCUR_HOME=<dir> to relocate all four path kinds under one root.

• Use per-kind env vars only when a specific kind must diverge: CONCUR_CONFIG_DIR, CONCUR_DATA_DIR, CONCUR_STATE_DIR, CONCUR_CACHE_DIR.

• Resolution order is per-kind env var, --home, CONCUR_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 concur-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:

  jsonCopy code
  {
    "mcpServers": {
      "concur": {
        "command": "concur-pp-mcp",
        "env": {
          "CONCUR_HOME": "/srv/concur"
        }
      }
    }
  }
  

Fleet precedence: an inherited per-kind env var overrides an explicit --home for that kind. Use CONCUR_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 CONCUR_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:

bashCopy code
concur-pp-cli recall "<user's question>" --agent

The response envelope:

jsonCopy code
{
  "query": "...",
  "normalized": "<normalized form>",
  "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": ["<trial command>", "concur-pp-cli learnings confirm 12"] }
  ],
  "playbook": {
    "query_family": "...",
    "playbook": {
      "steps": [ { "cmd": "<command with {slot} substitution>", "purpose": "..." } ],
      "entity_slots": ["$ENTITY"],
      "expected_tool_calls": 3
    },
    "slots_resolved": { "$ENTITY": { "token": "<live token>", "canonical": "<canonical>" } },
    "notes": "<workarounds + gotchas for this query family>"
  },
  "notes": "<duplicate surface for non-playbook callers>"
}

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:

Copy code
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 <id>` only after the trial verified the behavior.
       Reject a wrong candidate with `learnings reject <id>`.
    -> 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 `concur-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 <id> prints the candidate's full payload before materializing it - check that the printed payload matches the behavior you verified. learnings reject <id> tombstones the derivation signature so the same candidate does not resurface. The envelope carries only the few candidates worth acting on now; concur-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:<canonical> (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 concur-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 teaches do not fragment the store. Fire it after assembling your user-facing response but BEFORE emitting it, with a shell & so the call returns immediately:

bashCopy code
concur-pp-cli teach --query "<user's question>" --resource-type <type> --resource <id1> --resource <id2>
# (append shell `&` to background it)

Silent on success. Errors only land in teach.log under the resolved state dir. Teach the most specific resource - if the user asked a broad question and you walked through parent records to find the specific answer, teach the leaf id, not the parent. The CLI uses seeded entity_lookups for cross-alias resolution at recall time, so a teach under one alias (e.g., "Niners") satisfies future queries under another alias (e.g., "49ers", "San Francisco") automatically.

PII rule: teach the structural question with identifiers stripped - never include names, emails, phone numbers, account ids, or other personal identifiers in taught queries or notes. The CLI scans teach queries for obvious email/phone shapes and warns, but does not block; strip before teaching rather than relying on the warning.

Step 5: playbooks - optional flags, automatic synthesis

You do not need to decide whether a session "deserves" a playbook: a teach on a family without one auto-synthesizes a playbook_candidate from the session's journal, and the next session judges it via confirm/reject. Attach explicit playbook flags only when you already hold choreography worth recording verbatim - workarounds the CLI didn't surface (silently-dropped flags, undocumented params, pagination tricks, payload gotchas). Prefer the integrated one-call form - record the resource learning and the playbook in the same teach invocation:

bashCopy code
# Common case: record both the resource learning AND the playbook in one call.
concur-pp-cli teach \
  --query "<user's question>" \
  --resource <id> \
  --playbook-file ~/playbooks/<shape>.json \
  --playbook-notes-file ~/playbooks/<shape>-notes.md
# (append shell `&` to background it)

# Alternate: playbook-only (no resource to record alongside).
concur-pp-cli teach-playbook \
  --query "<user's question>" \
  --playbook-file ~/playbooks/<shape>.json \
  --notes-file ~/playbooks/<shape>-notes.md

Playbook files are JSON with steps, entity_slots, expected_tool_calls. Notes files are markdown carrying the gotchas verbatim. File-free callers (MCP-only agents) pass the same content inline: --playbook-json and --playbook-notes on the integrated teach form, --playbook-json and --notes on teach-playbook. On the integrated teach form, the playbook flags are optional - omit them entirely for a resource-only teach. On the standalone teach-playbook form, at least one of the playbook and notes flags must be set; both empty is rejected. Playbooks are keyed on the structural query family (entities stripped) so a recipe taught from one entity-shaped query applies to every other query of the same shape, with slots_resolved binding the live query's canonical at recall time.

When you DO find a playbook on a future recall, treat it as ground truth: replay the steps with slots_resolved substitutions, skip the discovery that the choreography already documents, and read notes before any step.

Step 6: playbook amend & when your debug response identifies a correction

If your debug-protocol response identifies a concrete correction the notes or playbook should know — a workaround, an undocumented endpoint shape, a stale field name, observed schema drift, an empty-payload fallback — fire playbook amend BEFORE emitting your user-facing response. Same fire-and-forget posture as teach.

bashCopy code
concur-pp-cli playbook amend \
  --query "<exact recall query string>" \
  --add-note "<your concrete correction>"
# (append shell `&` to background it)

What counts as worth amending: a behavior you OBSERVED this session that future-you would benefit from knowing. Examples worth amending:

• A workaround for a CLI surface that silently drops or misorders a flag.
• An undocumented endpoint shape (response wrapped in {meta, results}, payload nested two levels deeper than the docs claim).
• Observed schema drift (a field renamed, an index that shifted between seasons, a category label that the API now returns lower-cased).

What does NOT belong in notes:

• The year-specific or entity-specific answer to the user's question. That's the response, not a learning.
• Per-team / per-athlete / per-row data the playbook already retrieves at runtime.
• Statements that paraphrase what the existing notes already say.

The amend command appends to the family's existing notes with a timestamped marker ([amend YYYY-MM-DDTHH:MMZ]: <text>). Multiple amends accumulate; the audit trail is visible. If no playbook exists yet for the family, amend creates a notes-only one (so cold-start corrections still land).

PII discipline for amend notes

playbook amend notes are designed to potentially flow upstream as shared knowledge in future versions of the Printing Press. Keep them clean of user-identifying content so the upstream-contribution path stays open without retroactive scrubbing:

• Do NOT embed paths to user filesystems, personal API keys or tokens, user email addresses, user GitHub handles, or specific query histories tied to a single user.
• Acceptable: endpoint shapes, undocumented field names, API gotchas, observed schema drift, workarounds for CLI surfaces, generalizable pagination or retry tactics.

If a correction is only meaningful with user-specific context, it belongs in a personal note, not in the playbook amend.

Measuring the loop

concur-pp-cli learnings stats reports recall hit rate, teach-to-reuse, playbook resolution rate, and candidate confirm/reject counts from the local learn_events table. Rates are null until they have a denominator; everything stays on this machine. Use it to check whether the loop is earning its keep for this CLI.

Disabling learning

• --no-learn on a single command short-circuits both recall and the teach write path. Use for deterministic agent flows or tests that must not be affected by accumulated learnings.
• CONCUR_NO_LEARN=true in the environment globally disables the pipeline.

Agent Feedback

When you (or the agent) notice something off about this CLI, record it:

Copy code
concur-pp-cli feedback "the --since flag is inclusive but docs say exclusive"
concur-pp-cli feedback --stdin < notes.txt
concur-pp-cli feedback list --json --limit 10

Entries are stored locally as feedback.jsonl under the resolved data dir. They are never POSTed unless CONCUR_FEEDBACK_ENDPOINT is set AND either --send is passed or CONCUR_FEEDBACK_AUTO_SEND=true. Default behavior is local-only.

Write what surprised you, not a bug report. Short, specific, one line: that is the part that compounds.

Output Delivery

Every command accepts --deliver <sink>. The output goes to the named sink in addition to (or instead of) stdout, so agents can route command results without hand-piping. Three sinks are supported:

Unknown schemes are refused with a structured error naming the supported set. Webhook failures return non-zero and log the URL + HTTP status on stderr.

Named Profiles

A profile is a saved set of flag values, reused across invocations. Use it when a scheduled or recurring agent reuses the same saved flags while providing different input each run.

Copy code
concur-pp-cli profile save briefing --json
concur-pp-cli --profile briefing payment-types
concur-pp-cli profile list --json
concur-pp-cli profile show briefing
concur-pp-cli profile delete briefing --yes

Explicit flags always win over profile values; profile values win over defaults. agent-context lists all available profiles under available_profiles so introspecting agents discover them at runtime.

Async Jobs

For endpoints that submit long-running work, the generator detects the submit-then-poll pattern (a job_id/task_id/operation_id field in the response plus a sibling status endpoint) and wires up three extra flags on the submitting command:

Use async submission without --wait when you want to fire-and-forget; use --wait when you want one command to return the finished artifact.

Exit Codes

Argument Parsing

Parse $ARGUMENTS:

1. Empty, help, or --help → show concur-pp-cli --help output
2. Starts with install → ends with mcp → MCP installation; otherwise → see Prerequisites above
3. Anything else → Direct Use (execute as CLI command with --agent)

MCP Server Installation

1. Install the MCP server:

   bashCopy code
   go install github.com/mvanhorn/printing-press-library/library/productivity/concur/cmd/concur-pp-mcp@latest
   

2. Register with Claude Code:

   bashCopy code
   claude mcp add concur-pp-mcp -- concur-pp-mcp
   

3. Verify: claude mcp list

Direct Use

1. Check if installed: which concur-pp-cli If not found, offer to install (see Prerequisites at the top of this skill).
2. Match the user query to the best command from the Unique Capabilities and Command Reference above.
3. Execute with the --agent flag:

   bashCopy code
   concur-pp-cli <command> [subcommand] [args] --agent
   

4. If ambiguous, drill into subcommand help: concur-pp-cli <command> --help.