---
name: pp-nepra
description: "Pakistan's electricity regulator publishes the country's only plant-level power data as Excel-exported HTML that reads as empty and PDFs nobody extracts — this turns it into a queryable panel and names every gap. Trigger phrases: `pakistan power plant generation`, `nepra generation data`, `disco t&d losses`, `saifi saidi pakistan`, `pakistan electricity capacity by fuel`, `use nepra`, `run nepra`."
author: "qazmataz"
license: "Apache-2.0"
argument-hint: "<command> [args] | install cli|mcp"
allowed-tools: "Read Bash"
metadata:
  openclaw:
    requires:
      bins:
        - nepra-pp-cli
    install:
      - kind: go
        bins: [nepra-pp-cli]
        module: github.com/mvanhorn/printing-press-library/library/other/nepra/cmd/nepra-pp-cli
---
<!-- GENERATED FILE — DO NOT EDIT.
     This file is a verbatim mirror of library/other/nepra/SKILL.md,
     regenerated post-merge by tools/generate-skills/. Hand-edits here are
     silently overwritten on the next regen. Edit the library/ source instead.
     See the repository agent guide, section "Generated artifacts: registry.json, cli-skills/". -->

# NEPRA — Printing Press CLI

## Prerequisites: Install the CLI

This skill drives the `nepra-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 `$HOME/.local/bin` on macOS/Linux and `%LOCALAPPDATA%\Programs\PrintingPress\bin` on Windows:
   ```bash
   npx -y @mvanhorn/printing-press-library install nepra --cli-only
   ```
2. Verify: `nepra-pp-cli --version`
3. Ensure the reported install directory is on `$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/other/nepra/cmd/nepra-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.

NEPRA is the sole public source for monthly plant-by-plant generation and per-distribution-company reliability in Pakistan, and there is no API, no dataset and no package for any of it. This CLI extracts those releases live from the published documents, keeps unreported months distinct from measured zeros, follows plants across years despite renamed rows and re-sorted row numbers, and surfaces the places the regulator's own documents disagree with themselves instead of quietly picking one.

## When to Use This CLI

Reach for this CLI for any question about Pakistani electricity generation at plant granularity, distribution-company reliability against regulatory targets, licensed capacity, or fuel cost adjustment history. It is strongest on questions that need a time series or a cross-entity join that no single published page answers — a plant's seven-year load factor, a listed operator's consolidated fleet, or which distribution company's loss gap widened most. It is also the right tool when you need to know whether a figure is contested, because it records where the regulator's own documents disagree.

## Anti-triggers

Do not use this CLI for:
- Do not use this to audit or recompute a household or business electricity bill — the governing consumer tariff schedules are published as scanned images with no extractable rates, so any figure would be invented.
- Do not use this for net-metering or prosumer payback simulation; that is a modelling product and this CLI carries only published regulator data.
- Do not use this for load-shedding schedules or outage timetables; those are distribution-company publications, not regulator data, and are not carried here.
- Do not use this for consumer-end applicable tariff rates or fuel adjustment after the published consolidated series ends — the later determinations exist only as image-only or character-corrupted documents and the CLI refuses rather than guessing.
- Do not use this for circular debt, national sales or consumer-count aggregates; those live only inside very large State of Industry PDFs, one of which has no text layer at all.

## Unique Capabilities

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

### The panel nobody else has
- **`gen`** — Pull every power plant's monthly generation and utilisation for a fiscal year as rows you can pipe, with unreported months kept distinct from real zeros.

  _Reach for this when you need the whole fiscal year at plant granularity; it is the only machine-readable form of this table that exists._

  ```bash
  nepra-pp-cli gen --fy 2023-24 --agent
  ```
- **`fleet`** — Follow one plant or a listed operator's whole fleet across the fiscal years the crosswalk has observed, through renames and re-sorted row numbers.

  _Use this for any question spanning more than one fiscal year, and for mapping plants to listed operators. meta reports observed_fys against published_fys: the embedded crosswalk does not cover every published year, and the gap is stated rather than hidden._

  ```bash
  nepra-pp-cli fleet --parent HUBC --agent
  ```
- **`capacity`** — Installed and dependable capacity broken out by plant status and system, so each headline megawatt figure is labelled and derivable.

  _Use this when a capacity number has to reconcile; it shows which definition produces which total._

  ```bash
  nepra-pp-cli capacity --as-of 2024-06-30 --by status --agent
  ```

### Trust the number or don't
- **`conflicts`** — List every place NEPRA's own published figures disagree — two sources on one key, an unexplained step inside a single series, or a row that fails its own arithmetic — showing both values and both citations without picking a winner.

  _Reach for this before quoting any NEPRA figure. `--kind conflict` is the two-source case; `--kind break` is a step inside one series (those entries carry same_key=false, because there is no second source to cite); `--kind arithmetic` is a row that contradicts itself. Default is all three._

  ```bash
  nepra-pp-cli conflicts --surface per --agent
  ```
- **`disco`** — Per-distribution-company transmission losses, recovery, interruption frequency and duration against the regulator's own targets, across report years.

  _Use this for any per-DISCO performance question; the national averages everyone quotes hide the spread._

  ```bash
  nepra-pp-cli disco --metric saidi --fy 2024-25 --agent
  ```
- **`verify`** — Assert every published NEPRA surface against status, byte floor, content checksum, in-document year token and the frozen column fingerprint, exiting non-zero when stale or drifted.

  _Use this before trusting any extract. It is a FETCH gate, not a cache gate — nothing is stored locally — and it catches the decoys this site serves under an HTTP 200._

  ```bash
  nepra-pp-cli verify --fy 2023-24 --agent
  ```

### Find the document
- **`events`** — A dated feed of tariff determinations and orders by company, distribution company or docket, with the verbatim document link.

  _Use this to find which determination applies to a company and when; it returns document identity, not rates._

  ```bash
  nepra-pp-cli events --surface ipp-thermal --docket TRF-71 --agent
  ```
- **`licence`** — The licence register with gross capacity, plant type, fuel and the modification trail, searchable by name or fuel.

  _Reach for this for licensed-capacity and plant-type questions. Search by --search or --fuel; lookup by PSX ticker is refused with exit 2 because the crosswalk resolves only 28 of 335 register names, and fleet --parent is the right instrument for an operator's generation panel._

  ```bash
  nepra-pp-cli licence --fuel Coal --agent
  ```
- **`fca`** — The monthly requested-versus-allowed fuel cost adjustment series with cumulative disallowance, bounded to the window the source actually covers.

  _Use this to see how much of a requested fuel adjustment the regulator actually allowed, month by month._

  ```bash
  nepra-pp-cli fca --entity cppag --cumulative --agent
  ```
- **`sources`** — Catalogue every reachable published document with size, coverage window, text-layer flag and reachability state, and diff it against the shipped manifest.

  _Use this to discover what exists before fetching, and to tell an unreachable document apart from one with no data._

  ```bash
  nepra-pp-cli sources --kind per --agent
  ```

## Command Reference

**coverage** — The coverage ledger: what this CLI serves, from where, and how old

- `nepra-pp-cli coverage` — The coverage ledger: all 104 surfaces this CLI can serve, joined from its four in-code catalogues, with the command that answers for each and the age of every measurement. MAKES NO REQUEST. `--provenance` declares what a fetch of each surface records and what it deliberately does not; `--check-stale` exits 6 when ANY surface is past the window, which is what makes it usable as a cron gate.

**events** — Tariff determinations and Authority orders

- `nepra-pp-cli events` — The determinations feed. REQUIRES gzip: without Accept-Encoding this page silently truncates under a 30 s timeout and returns a partial list under an HTTP 200.

**fca** — Fuel cost adjustment: monthly requested versus allowed, CPPA-G and K-Electric

- `nepra-pp-cli fca` — The consolidated FCA table. Coverage is hard-bounded Jul-2018 to Jun-2022; negatives are written in accounting parentheses.

**generation** — Plant-level installed and dependable capacity with monthly generation and utilisation, per fiscal year

- `nepra-pp-cli generation index` — The site's own generation index. INCOMPLETE — links 5 of the 7 reachable years; never use as the sole enumerator.
- `nepra-pp-cli generation year` — Fetch one fiscal year's generation workbook payload. windows-1252 encoded; the HTTP header declares no charset.
- `nepra-pp-cli generation plants` — One row per plant for a fiscal year: installed and dependable capacity, technology, fuel and status, filterable with `--technology` and `--fuel`. Capacity is INDEPENDENT of the monthly block in both directions, and the `*_state` columns carry that: a delicensed plant keeps its published capacity, while eleven FY2017-18 plants publish no capacity number at all. NO capacity factor or utilisation is emitted at this or any grain — the source publishes no hours-per-month denominator.
- `nepra-pp-cli generation monthly` — Twelve rows per plant for a fiscal year, the monthly GWh and `% age` cells, same `--technology` / `--fuel` filters. Unfiltered it emits exactly what `gen --fy <year>` emits, from the same builder. A blank NEVER becomes a zero.

**hydel** — Hydel generation and dependable capacity

- `nepra-pp-cli hydel` — Hydel data. Dependable capacities are pinned upstream to 'July 2018 - June 2022'.

**licence** — Generation licence register: gross capacity, licence number, plant type, fuel, modification trail

- `nepra-pp-cli licence` — The licence register accordion. Key lookup must be typo-aliased: the page misspells its own column as 'Gross Capacityy', hiding the largest entry.

**quarterly** — Quarterly data for ex-WAPDA DISCOs and K-Electric

- `nepra-pp-cli quarterly` — Quarterly index. FROZEN upstream at 'April 2022- June 2022 (4th Quarter)' with 'Not Yet Issued' against every DISCO; nothing on the page says it is stale.

**reliability** — DISCO Performance Evaluation Reports: T&D losses against allowed target, recovery, SAIFI, SAIDI, complaints, safety

- `nepra-pp-cli reliability <path>` — Fetch one Performance Evaluation Report PDF. Filenames are unconstructible across years — always scrape from sources, never build the path.

**sir** — State of Industry Report PDFs

- `nepra-pp-cli sir <year>` — One State of Industry Report, returned as PDF bytes. Very large: 2025 is ~316 MB.

**sro** — SRO notifications

- `nepra-pp-cli sro` — SRO notification index.

**tariff** — Per-distribution-company notified tariff pages

- `nepra-pp-cli tariff <disco>` — One DISCO's tariff page. UTF-8 on this surface, unlike the generation workbooks. Rate schedules themselves are image-only PDFs and are NOT extractable.


### Finding the right command

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

```bash
nepra-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. `--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

### One operator's fleet across every published year

```bash
nepra-pp-cli fleet --parent HUBC --agent
```

Follows a listed operator's plants through renames and re-sorted rows, narrowing the payload to the four fields an availability model needs.

### Which distribution company misses its loss target by most

```bash
nepra-pp-cli disco --metric tnd --fy 2024-25 --agent
```

Returns actual reported losses alongside the level allowed in tariff and the breach, per company, which no published page tabulates together.

### Check a figure is not contested before quoting it

```bash
nepra-pp-cli conflicts --surface per --agent
```

Lists same-key disagreements between two published documents with both citations, so a quoted number can be defended or avoided.

### See which fiscal years and surfaces actually exist upstream

```bash
nepra-pp-cli sources --diff --agent
```

Diffs the live catalogue against the shipped manifest, separating genuinely unreachable documents from ones with no data.

### Gate a published surface on bytes, decoys and the column fingerprint

```bash
nepra-pp-cli verify --fy 2023-24 --check-stale --agent
```

Emits a citable checksum manifest and exits non-zero if any artifact drifted or aged past its staleness window.

## Auth Setup

No authentication required.

Run `nepra-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 (assertions, summaries and warnings 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
  nepra-pp-cli events --agent
  ```
- **Previewable** — `--dry-run` shows the request without sending
- **Live-only** — every data command reads the published documents on each invocation, or computes from an embedded catalogue. There is no sync command and no local NEPRA store; `--data-source local` is refused rather than answered emptily
- **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": <data>
}
```

Parse `.results` for data and `.meta.source` to know where it came from. `meta.source` is NOT just
live-or-local on this CLI — there is no local NEPRA store — and it takes four values:

| `meta.source` | Meaning | Commands |
|---|---|---|
| `live` | read from nepra.org.pk on this invocation | gen, generation plants/monthly, capacity, disco, events, licence, fca, verify, the promoted resources |
| `computed` | derived from an embedded crosswalk, no request | fleet |
| `ledger` | a shipped in-code registry, no request | conflicts without `--recompute` |
| `catalogue` | an in-code catalogue, no request | coverage, and the no-selector branch of gen/sources/events/licence |

TWO COMMANDS DO NOT USE THE TWO-KEY ENVELOPE, so an agent reading only `.results` loses data:

- `disco` emits a THIRD top-level key, `weighted_average`, carrying NEPRA's own published W.Av row
  as a separate series. It is deliberately not mixed into `.results`, because a published average
  is not another entity.
- `verify --manifest` emits no `meta`/`results` pair at all: it prints the frozen expectation
  manifest as its own document (`manifest_as_of`, `column_fingerprint`, `logical_columns`,
  `surfaces`) and makes no request.

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 `NEPRA_HOME=<dir>` to relocate all four path kinds under one root.
- Use per-kind env vars only when a specific kind must diverge: `NEPRA_CONFIG_DIR`, `NEPRA_DATA_DIR`, `NEPRA_STATE_DIR`, `NEPRA_CACHE_DIR`.
- Resolution order is per-kind env var, `--home`, `NEPRA_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 `data.db`, which holds only the Printing Press learn and playbook tables — there is no NEPRA data in it. `state` contains persisted queries, jobs, and `teach.log`. `cache` contains regenerable HTTP/cache files. THIS CLI STORES NO CREDENTIAL OF ANY KIND: there is no `credentials.toml`, no cookie jar and no auth sidecar, because NEPRA needs none.
- Run `nepra-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": {
      "nepra": {
        "command": "nepra-pp-mcp",
        "env": {
          "NEPRA_HOME": "/srv/nepra"
        }
      }
    }
  }
  ```

Fleet precedence: an inherited per-kind env var overrides an explicit `--home` for that kind. Use `NEPRA_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 `NEPRA_HOME`, or `doctor` will not find the learn/playbook database 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
nepra-pp-cli recall "<user's question>" --agent
```

The response envelope:

```json
{
  "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>", "nepra-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:

```
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 `nepra-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; `nepra-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.
- 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:

```bash
nepra-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:

```bash
# Common case: record both the resource learning AND the playbook in one call.
nepra-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).
nepra-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`.

```bash
nepra-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

`nepra-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.
- `NEPRA_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:

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

Entries are stored locally as `feedback.jsonl` under the resolved data dir. They are never POSTed unless `NEPRA_FEEDBACK_ENDPOINT` is set AND either `--send` is passed or `NEPRA_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:

| Sink | Effect |
|------|--------|
| `stdout` | Default; write to stdout only |
| `file:<path>` | Atomically write output to `<path>` (tmp + rename). Binary-response commands write decoded payload bytes (not the base64 JSON envelope) and print a small JSON receipt on stdout; `--json`/`--csv` do not refuse when this sink is set. |
| `webhook:<url>` | POST the output body to the URL (`application/json`) |

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.

```
nepra-pp-cli profile save briefing --json
nepra-pp-cli --profile briefing events
nepra-pp-cli profile list --json
nepra-pp-cli profile show briefing
nepra-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.

## Exit Codes

| Code | Meaning |
|------|---------|
| 0 | Success |
| 1 | A completeness assertion failed under `--strict` (gen, disco, verify, generation plants/monthly) |
| 2 | Usage error (wrong arguments), including a deliberate refusal such as `licence --ticker` |
| 3 | NEPRA does not publish what was asked for at the path this build knows |
| 5 | API or parse error (upstream issue, or a document that did not parse) |
| 6 | A staleness or fetch gate refused (`verify`, `coverage --check-stale`) |
| 7 | Rate limited (wait and retry) |
| 10 | Config error |

There is no exit 4: this CLI needs no credential, so it can never raise an auth error.

## Argument Parsing

Parse `$ARGUMENTS`:

1. **Empty, `help`, or `--help`** → show `nepra-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:
   ```bash
   go install github.com/mvanhorn/printing-press-library/library/other/nepra/cmd/nepra-pp-mcp@latest
   ```
2. Register with Claude Code:
   ```bash
   claude mcp add nepra-pp-mcp -- nepra-pp-mcp
   ```
3. Verify: `claude mcp list`

## Direct Use

1. Check if installed: `which nepra-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:
   ```bash
   nepra-pp-cli <command> [subcommand] [args] --agent
   ```
4. If ambiguous, drill into subcommand help: `nepra-pp-cli <command> --help`.