Planning 1on1 Annual Digest

---
name: planning-1on1-annual-digest
sheetId: "0.04"
description: >-
  Turn ONE teammate's shared 1:1 Doc (weekly WB entries), their LOB plan Doc
  and the tracker rows they own into a zh-TW 「年度計畫＿<name>」 Google Doc in
  the planning hub folder — six fixed sections (過去一年成果總覽 · 策略透鏡
  C1–C4 · 復盤 放大／收割／停止或改造 · 年度計劃 3–5 主軸 · OKR · 接下來) with
  every claim tagged to its WB date, plus a 12-line slide summary block for
  that person's session one-pager. Trigger on /planning-1on1-annual-digest or
  when Peter says "幫 <name> 整理年度計畫", "從 1:1 文件整理 <name> 的年度回顧",
  "把 <name> 的 1:1 筆記變成年度計畫", "做 年度計畫＿<name>", "annual digest
  for <name>", or hands over a 1:1 Doc ID before a planning session. Distinct
  from planning-prework-pack (ALL owners → session workbook + deck; it CONSUMES
  this skill's slide block), from planning-evidence-pack (company-level numbers
  scoreboard), from planning-session-synth (digests the SESSION transcript, not
  1:1 docs), and from project-status-update (the course-project WEEKLY — it
  owns bare 「週報」/「weekly report」). Not a performance review: it never
  grades the person, never invents metrics, never writes into the 1:1 Doc.
category: strategy
project: planning-1on1-annual-digest
platform: claude
status: Done
author: Peter Tu
input: "One person's name (→ 1:1 Doc via sources §B, or a Doc ID) · optional cycle H1/H2/YE (default YE) · optional window (default last 12 months) · optional LOB / plan Doc override"
process: "Step 0 resolve cycle + sources → parse WB entries into a dated ledger → read LOB plan Doc + owned tracker rows → build 一~六 (timeline · C1–C4 lens · 放大/收割/停止 · 主軸 · OKR · 接下來) → show outline + 待補 list → create Doc in hub → 12-line slide block"
output: "「年度計畫＿<name>」 Doc in the planning hub folder (zh-TW, six exact sections, WB-dated claims, 待定/（待補） where no source) + a 12-line slide block in chat + a wrote / did-not-touch report"
synergy:
  - "planning-prework-pack"
  - "planning-evidence-pack"
  - "planning-suite-reconciler"
  - "project-status-update"
house-style: bound

---

# Planning 1on1 Annual Digest

```bash
npx skills add https://github.com/peter-tu-zynkr/zynkr-skill-builder --skill planning-1on1-annual-digest
```

Before a planning session every owner needs a one-page answer to "what did I
ship, what does the strategy want from my line, what do I commit to next?" — and
the raw material sits in fifty weekly 1:1 entries, a LOB plan Doc and a few
tracker rows. This skill digests those three sources for ONE person into the
zh-TW 「年度計畫＿<name>」 Doc the team already uses (exemplar 「年度計畫＿Peggy」,
sources §A): a WB-dated timeline, the C1–C4 lens, a 放大／收割／停止或改造 retro,
3–5 ordered 主軸 with 關鍵路徑, an OKR block and three things to do this month.
Unlike "summarise my notes", every claim carries its WB date, every number
carries its source or is written 待定, the strategy lens is quoted from the plan
Docs rather than opined, and the 1:1 Doc itself is never touched.

---

## How this differs from its neighbours

- **planning-prework-pack** (0.03) — the session-level aggregator: ALL owners →
  workbook tabs + one-pagers + deck request. It CONSUMES the 12-line slide block
  this skill prints. This skill goes deep on ONE person; run it once per owner.
- **planning-evidence-pack** (0.05) — company-level "looking back in numbers"
  scoreboard. This skill borrows a proof number from it when one exists but
  never computes company metrics itself.
- **planning-session-synth** (0.06) — digests the SESSION transcript + whiteboard
  into tracker tabs; different input, different output.
- **project-status-update** (3.09) — the course-project WEEKLY; it owns every
  bare 「週報」/「weekly report」 trigger and all scheduled runs. This skill is
  annual, per person, Peter-invoked.

## Fixed facts (read the references first)

- **IDs never live in this file.** Hub folder, Main Tracker + SOR tab, per-LOB
  plan Docs, integrated plan Doc, OKR & KPI Tracker and the exemplar
  「年度計畫＿Peggy」 are in `./references/planning-sources.md` §A; the per-person
  1:1 Docs are in §B. A new cycle edits the sources file, not this skill.
- **Google account + MCP server** for every Docs/Drive/Sheets call: as stated in
  the sources file header (`google-workspace`; override only if the user says so).
- **Doc skeleton + slide block** = `./references/annual-plan-template.md`
  (section titles verbatim; plain headings + bullets; evidence-tag rule).
- **1:1 Doc parsing contract** = `./references/wb-entry-parsing.md`
  (`### WB（YYYY/M/D）` entries newest-first; the real label variants — 完成的事情／
  需完成的事情／本週項目／待處理／確認事項／Talking Points／Notes … — mapped in its §2;
  the per-entry 「Peter」 sub-block and 我的感覺 score handled by its §2a; the bottom
  「Read me」 template block is ignored).
- **Retro filter** 放大／收割／停止或改造 = pack §6 (last paragraph) · **C1–C4 +
  sequencing** = pack §4 · **taxonomy** = pack §2 · **tracker layout + status
  vocabulary** = pack §6 + §3 · **never-do list** = pack §9.
- **Doc creation path** — write the filled template to a local `.md`, then
  `import_to_google_doc(file_name, file_path, folder_id=<hub folder>)`: it accepts
  the folder directly and converts `#`/`##` headings + `-` lists to native Docs
  styles (tables / checkboxes do NOT survive — hence the no-table rule). Fallback
  only if the import errors: `create_doc(content=…)` (plain text, lands in My Drive
  root) → `update_drive_file(add_parents=<hub folder>)`. `create_drive_file` inside
  a folder returns HTTP 400. Either way the new Doc is **private to the account
  until shared** — sharing is a separate, confirmed step (Step 8).

## Hard rules

1. **Every claim quotes its evidence.** Sections 一 and 三 tag each bullet
   `（WB YYYY/M/D）` or `（Tracker #N.NN）`; strategy lines in 二 tag the plan Doc.
   An untagged claim is deleted, not kept.
2. **No invented metrics.** Proof numbers come from a WB line (as the person
   wrote it), a tracker cell, the OKR tracker, a plan Doc or the user — else
   `（待補）` for evidence and `待定` for a KR target (pack §9).
3. **The 1:1 Doc is read-only — so are the tracker, the OKR tracker and every
   plan Doc.** The only writes are ONE new Doc in the hub folder and the chat
   output. Never append to, comment on, or reformat the 1:1 Doc.
4. **Ignore the Read me / template block** at the bottom of the 1:1 Doc; a
   template line is never evidence. **Attribute the two voices**: bullets under
   an entry's 「Peter」 sub-block (and the template's "What's on my mind (Peter)"
   field) are the manager's asks — never the person's 完成／待完成／Action, never
   cited in the Doc (parsing file §2a).
5. **Never assume the roster.** Person → 1:1 Doc comes from sources §B or the
   user; an unlisted person ⇒ ask for the Doc ID, never search Drive by name and
   guess. 協作主管 comes from the tracker or the user.
6. **Six sections, exact titles, zh-TW, plain headings + bullets.** No tables,
   no checkboxes, no extra sections — the slide block lives in chat, not the Doc.
7. **Show before writing.** Print the outline + 待補／待定 list + coverage line;
   create the Doc after one confirmation (「直接寫」/`--write` up front
   pre-authorises). Never overwrite an existing Doc (pack §8).
8. **Not a review.** No grading language, no comparison to teammates. The
   「我的感覺：__N__分」 morale score is **manager-private**: one line in the chat
   report for Peter only — never in the Doc, never in the slide block. Slide
   `Top risk` is a business risk (delivery／capacity／dependency／market → C-n),
   not a morale reading.

---

## Workflow

### Step 0 — Resolve cycle + sources

Read `./references/planning-sources.md`. Take from the user: `cycle` (`H1` / `H2`
/ `YE`; default `YE` — a half-year run is legal), the window (default the 12
months ending on the session date or today), and any ID overrides (hub folder,
tracker, plan Doc, 1:1 Doc). Print one line before doing anything else:

```
Cycle: YE 2026 · window 2025-12-01 → 2026-11-30 · hub folder <id, sources §A> · tracker <id> (SOR tab per pack §6) · 1:1 Doc <id, sources §B or user>
```

Anything unresolved ⇒ stop and ask; never proceed on a guessed ID.

### Step 1 — Resolve the person and their line

- **Person** → 1:1 Doc ID (sources §B or user; hard rule 5).
- **Line (LOB)** → the L1 number(s) the person owns, derived from the tracker's
  負責人 column (Step 3) and confirmed with the user when they span two L1s; then
  the matching per-LOB plan Doc(s) from sources §A (6.0 is a shortcut — resolve
  it as the sources note says).
- **協作主管** → from the user or the tracker; else `（待補）`.
- **Existing Doc check** — `search_drive_files` Drive-wide by title
  (`name contains '年度計畫＿<name>'`; the hub holds shortcuts, so a prior Doc may
  live in a pillar folder). If found, title the new one `年度計畫＿<name>（<cycle> YYYY-MM）`
  and leave the old untouched (say so in Step 8). Never reuse a title.

### Step 2 — Parse the 1:1 Doc into an evidence ledger

`get_doc_as_markdown(comment_mode="none")` on the 1:1 Doc (Doc comments are usually
the manager's — never ledger them as the person's evidence), then apply
`./references/wb-entry-parsing.md`:
split on `WB（YYYY/M/D）` headings, stop at the Read me block, map each block
label to a ledger category via the parsing file's §2 table (完成的事情／已完成的部分
… → 完成; 需完成的事情／本週項目／待處理／確認事項／Talking Points／Notes … → 待完成),
and turn every bullet into a ledger row `date · category · verbatim line · tag ·
asset?`. Bullets under the entry's 「Peter」 sub-block go to a separate `主管` bucket
(the manager's asks — never the person's evidence, never counted); 我的感覺 scores go
to a manager-private note (parsing file §2a). Keep only rows inside the window
(older rows are context, never citations). Print the coverage line:

```
1:1 coverage: 41 WB entries · 2025/12/2 → 2026/11/25 · gaps ≥3 wks: 2026/1/13→2026/2/10 · 完成 rows 118 · Action rows 37 (29 closed by a later 完成)
```

Judgement calls (ambiguous dates, plan-sounding 完成 lines, alias merges) go on a
判讀 list printed in Step 8 — never silently resolved (pack §9).

### Step 3 — Read the strategy sources

- **Tracker rows they own** — `get_spreadsheet_info` to find the SOR tab
  (`<cycle> 專案項目`, pack §6), `read_sheet_values` on it, keep rows whose 負責人
  equals the person (協助者 rows noted separately). Normalise 狀態 to the pack §3
  vocabulary (trim; `完成`/`done`/`✓` → 完成 · `進行中`/`wip` → 進行中 ·
  blank/`未開始` → 未開始 · `放棄`); anything else stays `（狀態未知：<原值>）` and is
  flagged, never guessed. Skip L1 header rows (`1.0`, `2.0` …). Record `#`, 項目,
  Priority, 狀態, 開始／結束: 完成 rows are evidence for 一; P0/P1 未開始／進行中 rows
  are 主軸 candidates for 四.
- **LOB plan Doc(s)** — `get_doc_as_markdown`; pull the mandate for this line, the
  strategic pillars, and — from the dated Refresh addendum at the top (pack §8) —
  the current P0 list. Read the integrated plan Doc only for the 核心命題 and the
  C1–C4 fill when the LOB Doc is silent.
- **OKR & KPI Tracker** — read `OKRs` for KRs already on this person (Owner
  column) so 五 reuses committed numbers instead of minting new ones.
- **Optional extra evidence** the user hands over (Fireflies 1:1 recaps, an
  evidence-pack scoreboard) is tagged with its own source.

### Step 4 — Build 一、過去一年成果總覽

From the ledger's 完成 rows + tracker 完成 rows: group by `tag` (product / project
/ system), order groups by first appearance, and inside each group write month
bullets `YYYY/MM：<one-line outcome>（WB YYYY/M/D）` ascending; a deliverable
reported twice is one bullet with both dates. Then 「一起長出來的系統資產」 = every
`asset: yes` row (SOP · template · automation · tracker · Doc · course module)
with one clause on the dependency it removed or the ladder gap it filled. Numbers
stay as the person wrote them; a bullet that needs a proof number gets
`（待補：<the exact source needed>）`.

### Step 5 — Build 二、策略透鏡

Three blocks, all quoted, none opined: **核心命題** (one line from the integrated
or LOB plan Doc, tagged) · **四大限制 C1–C4** (pack §4 rows, filled from the plan
Docs' Refresh addendum; blank → `待定（計畫文件未明說）`) · **你這條線的任務** (the
LOB Doc mandate + this person's P0 items from Step 3, with Tracker #). Section
title reads `二、策略透鏡：<cycle> 在打什麼`.

### Step 6 — Build 三、復盤：過去成果 × 策略支柱

Sort every group from 一 (and each major asset) into exactly one of 放大 ／ 收割 ／
停止或改造 using the pack §6 retro filter, ending each item with the constraint it
hits (`— 打到 C2：…`). Signals: a 這周思考 line arguing for/against the work, a
tracker row marked 放棄, a plan-Doc pillar naming it, repeated 待完成 rows with no
closure (a 停止或改造 candidate, said neutrally). Every item keeps its WB tag; an
item absent from 一 cannot appear in 三.

### Step 7 — Build 四、年度計劃 · 五、年度目標（OKR）· 六、接下來

- **四** — 3–5 主軸 ordered by pack §4's sequencing rule (cheap compounding work
  first; capital-heavier convert/scale after, gated). Each names its constraint +
  pillar and one 關鍵路徑 line (milestone → milestone → milestone, `Tracker #` where
  the item exists). Candidates: Step 3's P0/P1 rows, the LOB Doc's forward
  section, the ledger's open 待完成／Action rows — never thin air.
- **五** — one O mapping back to 二's 核心命題; KR1–KR5 (KR5 optional). A KR ends
  with a number **and its source** (`— 12 場（Tracker #3.04）`) or `待定`. KRs
  already on this person in the OKR tracker are reused verbatim.
- **六** — describe how O/KRs would land in the OKR tracker (`OKRs` tab: Objective ·
  KR · Owner · Tracker #) **without writing there** — that write belongs to
  planning-suite-reconciler or a human — then 「這個月先動 3 件事」: exactly three,
  each pointing at a 主軸 or Tracker #, preferring items whose 開始 date is already
  inside this month.

Run the acceptance checks in `./references/annual-plan-template.md` §C, then print
the outline (section titles + first bullet per block), the 待補／待定 count with
locations and the 判讀 list, and ask for the go-ahead (hard rule 7).

### Step 8 — Create the Doc, print the slide block, report

**Doc** — fill `./references/annual-plan-template.md` §A (in-Doc H1
`<line>・年度計畫（<cycle>）` + header lines 主體／策略依據／協作主管／整理範圍 +
sections 一–六 verbatim; teammate-facing wording — no pack §／skill／tool names in
the body), save it as a local `.md` (`#` H1 · `##` 一–六 · `-` bullets), then:

```
## preferred: import (folder_id lands it in the hub directly; #/## → native headings)
mcp__google-workspace__import_to_google_doc(user_google_email="<account from sources header>",
  file_name="年度計畫＿<name>", file_path="<local .md>", source_format="md",
  folder_id="<hub folder id from sources §A>")
## fallback only if the import errors: plain text, then move
mcp__google-workspace__create_doc(user_google_email="<same>", title="年度計畫＿<name>",
  content="<filled template>")
mcp__google-workspace__update_drive_file(user_google_email="<same>",
  file_id="<doc id>", add_parents="<hub folder id>")
```

Tables and checkboxes do not survive the import — the template is headings +
bullets only. Read the Doc back once (`get_doc_as_markdown`) to confirm the six
titles survived and nothing from the 「Peter」 sub-block or 我的感覺 leaked in.

**Slide block** — print `./references/annual-plan-template.md` §B (twelve lines:
header · Top-3 delivered with WB dates and proof or `（待補）` · Top-3 goals with
numbers or `待定` · Top risk = a business risk → C-n, never the morale score ·
Source · 待補 count) for planning-prework-pack — this block is team-visible.

**Report** — end with (names come from the tracker's 負責人 column at run time;
the block below is illustrative):

```
已寫入：年度計畫＿王小明 — <doc url>（hub folder：<folder url>）· 未分享（目前僅 <account> 可見）
Slide block：見上（12 行，可貼給 planning-prework-pack）
待補／待定：5 處（一、×2 · 五、KR3 · KR4 · 二、C1）
判讀：日期推定 1 筆（WB 2026/5/6？）· 別名合併 1 組（客服 SOP／客服流程）· Peter 子區塊排除 6 條
私下備註（僅供主管）：我的感覺 4 週 ≤ 5 分（2026/3–4）— 未寫入 Doc／slide
未做：未動 1:1 Doc · 未動 tracker／OKR tracker／計畫 Doc · 未分享 Doc · 未寄任何信 · 未建 GM Knowledge Directory 目錄項（如需，人工加）
要分享給 王小明 嗎？（確認後 manage_drive_access grant → 該員 email，commenter，不寄通知；hub folder 既有權限不變）
```

**Share (only after a second confirmation)** — the Doc is private to the account
until shared. If Peter says yes, `manage_drive_access(file_id=<doc id>,
action="grant", share_with=<the person's work address, from the user>,
role="commenter", send_notification=false)` (or `role="writer"` if he says so) —
`send_notification=false` is mandatory: this skill never sends mail, so the
person is told by Peter, not by a Drive e-mail. Then append 「已分享：<name>
（commenter）」 to the report. No confirmation ⇒ leave it unshared and say so;
never share with anyone but the person named.

---

## Outputs

- **`年度計畫＿<name>` Google Doc** in the planning hub folder — zh-TW; H1
  `<line>・年度計畫（<cycle>）`; header lines 主體／策略依據／協作主管／整理範圍;
  sections 一、過去一年成果總覽 · 二、策略透鏡 · 三、復盤 · 四、年度計劃 ·
  五、年度目標（OKR）· 六、接下來, shaped exactly as `./references/annual-plan-template.md` §A.
- **12-line slide summary block** in chat — the per-owner input
  `planning-prework-pack` expects.
- **Run report** in chat: Doc URL, coverage line, 待補／待定 list, 判讀 list, and
  the explicit 未做 list.

## Reference files

- `./references/planning-knowledge-pack.md` — shared family pack (byte-identical
  across `planning-*`; do not edit here): §2 taxonomy · §3 priority + status ·
  §4 C1–C4 + sequencing · §6 tracker layout + retro filter · §8 versioning · §9 never-do.
- `./references/planning-sources.md` — shared live IDs (byte-identical): §A hub
  folder · tracker · plan Docs · OKR tracker · exemplar; §B per-person 1:1 Docs.
- `./references/annual-plan-template.md` — this skill's Doc skeleton (§A), the
  12-line slide block (§B) and the pre-write acceptance checks (§C).
- `./references/wb-entry-parsing.md` — how a 1:1 Doc becomes a dated evidence
  ledger: entry boundaries, field labels, ledger row shape, window + coverage,
  judgement calls to surface.

## Limitations

- One person per run; a team of six = six invocations (then
  `planning-prework-pack` aggregates). It never ranks people against each other.
- Evidence-bound: a thin 1:1 Doc yields a thin, honest Doc with many `（待補）` —
  never a padded one. Weeks with no WB entry are reported as gaps, not filled.
- Strategy content is quoted, not authored: if the plan Docs carry no C1–C4 fill
  or no mandate for the line, section 二 says 待定 and the report flags it.
- Reads the OKR tracker but never writes it; the 六 mapping is a hand-off to
  `planning-suite-reconciler` or a human.
- No tables/checkboxes in the Doc (they do not survive `import_to_google_doc`) —
  the exemplar's shape is reproduced with headings + bullets. No Google Chat,
  calendar counting or cron — extra evidence enters only when the user hands it
  over.
- The Doc is created unshared; sharing with the person is a separate confirmed
  step, and the manager-private 我的感覺 note never leaves the chat.

## House style

Writing style is **not owned by this file**. The house voice lives in two Google Docs under
`[@] 寫作指南` (`12DBdFz3SK22ie9im_ThFMI7IBRXsTZsV`), read at runtime:

- 《[2.0] Zynkr 通用風格指南 House Voice》 `10bOIQwRm9Pxwgct4hlwCwK_B4Pipai1HqBPZKzyRHSE` —
  the universal core, plus the addendum for this surface
- 《[3.2] 禁用詞清單 Forbidden Words》 `1N5sHLP4qzmmhpCGsi6KElxi1z0MFe4QZ0Q_35T10Uyg`

Read both before producing client- or reader-facing text, and scan the draft against 《[3.2]》
before handing it over. If Drive is unreachable, say so in the output rather than proceeding
unchecked. Never re-implement either list inside this file.