Przejdź do treści
Biblioteka skilli
Zarządzanie i produkt · Planowanie

pano

/pano

Prowadzi zadania projektu w tablicy Pano: planuje PBI, zapisuje weryfikację i przekazuje pracę do testów.

Repozytorium
pano ↗
Licencja
MIT
Ostatnia zmiana repozytorium
09.11.2026

Ocena i bezpieczeństwo

100/ 100Klasa: dobry
Format25 / 25
Frontmatter z nazwą i opisem, poprawna nazwa, opis do 1024 znaków, nazwa zgodna z folderem.
Odkrywalność25 / 25
Opis mówi, kiedy użyć skilla, ma sensowną długość i jest w 3. osobie.
Treść35 / 35
Co najmniej 150 słów i do 500 linii, nagłówki, listy kroków, przykłady albo kod.
Higiena15 / 15
Bez zaślepek typu TODO, bez ścieżek Windows, z licencją.
Ryzykobrak sygnałów w skanie
Pliki w paczce
3
Skrypty
2
Komentarze HTML
0
Adresy URL w SKILL.md
1
Przestarzałość

Brak sygnałów. Repozytorium zmienione 09.11.2026, bez odwołań do wycofanych modeli i API.

Licencja

MIT, otwarta: można używać i zmieniać, z zachowaniem informacji o autorze.

Co sprawdzić przed instalacją
  • 2 skrypty w paczce. Przejrzyj je przed instalacją, bo Claude może je uruchomić.

Skrypty: scripts/usage_from_claude.py, scripts/usage_from_codex.py

Co robi i kiedy

Prowadzi zadania projektu w tablicy Pano: planuje PBI, zapisuje weryfikację i przekazuje pracę do testów.

Oryginalny opis (EN)

Track project work in Pano. Use when asked to plan PBIs, update a board, record verification, or hand work to a person for testing.

Jak działa

  1. Connect
  2. Workflow
  3. Read usage
  4. Commits
  5. CLI examples
Sekcje z nagłówków SKILL.md, w oryginale · 52 linii, 716 słów
Podgląd SKILL.md
---
name: pano
description: Track project work in Pano. Use when asked to plan PBIs, update a board, record verification, or hand work to a person for testing.
---

# Pano

Pano tracks work; it does not run agents. Work within the requested project and scope.

## Connect

Use the available Pano MCP tools, or the `pano` CLI. The default server is `http://127.0.0.1:7777`. Read its `/agents.md` guide for current tool and API details. If MCP is unavailable, use the CLI or REST; do not claim an update succeeded without checking the result. In a source checkout, `bun run pano` runs the CLI.

Set `PANO_URL` if the server uses a different address. If the installation requires authentication, configure `PANO_TOKEN` in the client environment; do not put secrets in cards or output. Identify changes with `PANO_ACTOR` for the CLI, an MCP actor, or `X-Pano-Actor` for REST, using an identity such as `claude:po` or `codex`.

## Workflow

1. Read the board, project settings, Definition of Done, and relevant item before changing work. Create a PBI with a clear outcome and verifiable acceptance criteria when planning new work. Avoid duplicate items.
2. Assign work to the appropriate agent by tool name, such as `codex`, and move the item to In Progress. A PO coordinates and verifies delegated work when the user requests that role.
3. Record decisions, progress, and actual verification results in the item. Use MCP `add_record` with `kind: "check"` or REST for test results; CLI `task note` creates a note. Mark only criteria that have been verified, using their returned IDs.
4. After every run on an item, before moving it on, you must report usage: model, tokens, peak context size, and duration. This applies to every agent and tool, including a coordinating PO. Use MCP `record_usage`, REST `POST /items/{key}/usage`, or `pano task usage`. Never invent values; omit anything you cannot read. Record a run covering several items once, on one item, with `--note` (or `note`) naming all covered items.
5. If a person needs to verify the result, add scenarios with steps and expected outcomes, assign to `human`, and move the item to Test. Tell the user what to try. Otherwise, move it to Done after verification if project settings allow agents to close it.
6. Keep history. Cancel obsolete work with a reason rather than deleting it.

## Read usage

The bundled readers need only Python's standard library and print JSON:

- Claude Code logs: `~/.claude/projects/<project>/`. Run `python scripts/usage_from_claude.py <transcript.jsonl> [--since 2026-11-07T15:00:00Z]` from this skill folder. `--since` limits usage to the current run; repeated lines for one message count once.
- Codex logs: `~/.codex/sessions/`. Run `python scripts/usage_from_codex.py <rollout.jsonl>` or `python scripts/usage_from_codex.py --cwd <folder>`. Resumed sessions share a file; record a session covering several items once and name them in `--note`.
- Other tools: use the usage the tool reports, for example OpenCode's session result.

Send the known usage fields from the JSON, omitting nulls and reader metadata (`messages`, `file`, `cwd`). Cached reads are excluded from the readers' totals; peak context is separate from usage.

## Commits

Follow the target repository's `AGENTS.md` and its commit rules. In the Pano repository, the orchestrator writes commit titles in English with the card number and a clear outcome, for example `PANO-39: Simplified agent working instructions`. Keep technical terms in English, add 2 to 5 user-facing changes in the body, and review `git status` and `git diff --staged` before committing. Delegated agents do not commit or push.

## CLI examples

Replace `DEMO` with the requested project key and use the item key and criterion ID returned by Pano:

```sh
pano project list --json
pano board DEMO --json
pano task create DEMO "Add search" --body "Why this is needed, the scope and any decisions." --criteria "Search finds matching titles" --priority medium --json
pano task assign DEMO-1 codex
pano task move DEMO-1 in_progress
pano task show DEMO-1 --json
pano task check DEMO-1 <criterion-id>
```

Use `pano --help` for commands and flags. MCP `list_projects`, `get_board`, and `get_item` provide equivalent reads. To report known usage, for example, use `pano task usage DEMO-1 --model MODEL --input 12000 --output 3000 --duration 120000`. Duration is in milliseconds. Only include fields you know; omit unknown values. `--total` is the run's own total, while `--context` is peak context and is not added to token usage.

If asked to migrate a Markdown backlog, use `pano import backlog --project DEMO --json`. It imports `PBI-*.md`, creates a missing project, and skips source IDs already imported. It does not synchronize later Markdown edits.
Pliki w paczce (3)
  • SKILL.md 5 KB
  • scripts/usage_from_claude.py 3 KB
  • scripts/usage_from_codex.py 3 KB

Wersje

W bazie nie ma innych skilli o nazwie pano.

Migawka bazy z 13.11.2026 · wersja paczki 36af4c795d92 (skrót SHA-256 całej paczki)

Pobrania w ostatnich 30 dniach: 0 · ranking

Zgłoś problem z tym skillem: szkodliwa treść, błędny opis albo prośba autora o usunięcie.

Podobne skille

Brak podobnych skilli w bazie.

Uwagi i zgłoszenia

Wiadomość trafi do Bartosza Sokolińskiego. Adres e-mail jest potrzebny tylko do odpowiedzi. Szczegóły w polityce prywatności.