From dd7db9e662c947dc0985c2e42efc00222e43cc32 Mon Sep 17 00:00:00 2001 From: m3ta-chiron Date: Thu, 5 Nov 2026 13:24:38 +0100 Subject: [PATCH] feat(lead): Modus ich ersetzt meins im Lead-Radar MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit - Neuer Modus `ich` (Wochenselbstreview) über alle Projekte statt nur 🔒 Meins: Zusagen aus Sally, Termine, Abgeben, Fokus und Kennzahlen mit Verlauf - `references/meins.md` entfällt, ersetzt durch `references/ich.md`; Trigger und Invocation in SKILL.md angepasst - `tag`: neuer Block „Du heute“ mit eigenen fälligen und mehrfach verschobenen To-dos - `execute.md`: To-do anlegen nur im Modus `ich`, nur für eigene Zusagen - `lead-meeting-prep`: Meins-Items über `--projects` statt `--scope meins` laden --- skills/lead-meeting-prep/SKILL.md | 2 +- skills/lead-radar/SKILL.md | 19 +++-- skills/lead-radar/references/execute.md | 3 + skills/lead-radar/references/ich.md | 105 ++++++++++++++++++++++++ skills/lead-radar/references/meins.md | 61 -------------- skills/lead-radar/references/tag.md | 10 +++ 6 files changed, 129 insertions(+), 71 deletions(-) create mode 100644 skills/lead-radar/references/ich.md delete mode 100644 skills/lead-radar/references/meins.md diff --git a/skills/lead-meeting-prep/SKILL.md b/skills/lead-meeting-prep/SKILL.md index 4ee5464..44ec814 100644 --- a/skills/lead-meeting-prep/SKILL.md +++ b/skills/lead-meeting-prep/SKILL.md @@ -64,7 +64,7 @@ Run from `/home/sascha.koenig/p/PM`: ```bash python3 tools/it-cockpit/cockpit.py --json --scope runde --projects # Bezug projects of the round -python3 tools/it-cockpit/cockpit.py --json --scope meins # 🔒 Meins by list → the round's Meins lists +# add the 🔒 Meins project id (systems/basecamp.md) to --projects; filter its items by `list` for the round's Meins lists ``` `runde` returns per project `offen` (count), `ueberfaellig`, `diese_woche`, `still_30d`, `wartend_nachfassen`, each item with title, list, assignees, due, `still_tage`, description and URL. For a round whose Bezug is the whole portfolio (Weekly GF), use `--scope portfolio` instead of `runde`. Report anything under `errors` or `drift` in the prep footer. The cockpit takes 1–2 minutes; run both calls in parallel if the harness allows. diff --git a/skills/lead-radar/SKILL.md b/skills/lead-radar/SKILL.md index 19798fb..592c880 100644 --- a/skills/lead-radar/SKILL.md +++ b/skills/lead-radar/SKILL.md @@ -4,11 +4,12 @@ description: | Sascha's leadership radar: read the IT cockpit through one lens, turn the situation into numbered decision proposals, execute them in Basecamp/Zammad after one batch approval. Modes: tag (morning Lagebild, Zammad + 👥 IT), freitag (prepare the Friday close), - meins (weekly review of Sascha's own 🔒 Meins to-dos), person (prepare a 1:1), + ich (weekly self-review: Sascha's own to-dos across all projects, commitments from Sally, + delegation candidates, metrics against the last runs), person (prepare a 1:1), portfolio (monthly review of Sascha's projects). Also answers "wem delegiere ich X" via person mode. - Triggers: "Lagebild", "Freitags-Abschluss vorbereiten", "Wochenreview", "1:1 vorbereiten", - "Portfolio-Review". + Triggers: "Lagebild", "Freitags-Abschluss vorbereiten", "Wochenreview", "meine Woche", + "Selbstreview", "1:1 vorbereiten", "Portfolio-Review". compatibility: opencode, claude-code --- @@ -16,7 +17,7 @@ compatibility: opencode, claude-code Guiding question of every run: *What must Sascha decide now, and whom does he give it to?* One run = one mode → one Lagebild → one numbered proposal list → one batch approval → execution. -Invocation: `/lead-radar |portfolio>`. Without a mode, pick from the user's words via the trigger phrases above; still unclear → ask with the five modes as options. +Invocation: `/lead-radar |portfolio>`. Without a mode, pick from the user's words via the trigger phrases above; still unclear → ask with the five modes as options. ## Sources of truth @@ -56,7 +57,7 @@ From `/home/sascha.koenig/p/PM`: python3 tools/it-cockpit/cockpit.py --json --scope [--person ] ``` -A run takes 20–90 s. Save the JSON to the scratchpad and work from that file. Every output carries `drift` and `errors`; an entry under `errors` (e.g. Zammad without `ZAMMAD_TOKEN`) means that source is blind for this run: say so in one line at the top of the Lagebild and continue with the rest. Use live `basecamp` calls only for detail questions the cockpit does not answer (comments under an item, a card's history). +For `ich` add `--dry-run`: the state is written only at the very end of the run, after execution (see `references/ich.md`). A run takes 20–90 s. Save the JSON to the scratchpad and work from that file. Every output carries `drift` and `errors`; an entry under `errors` (e.g. Zammad without `ZAMMAD_TOKEN`) means that source is blind for this run: say so in one line at the top of the Lagebild and continue with the rest. Use live `basecamp` calls only for detail questions the cockpit does not answer (comments under an item, a card's history). **Done when** the JSON is on disk and you know which sources are blind. @@ -68,7 +69,7 @@ Open the mode's reference and follow it; it says which keys to read, how to judg |---|---|---| | `tag` | [references/tag.md](references/tag.md) | nothing | | `freitag` | [references/freitag.md](references/freitag.md) | nothing | -| `meins` | [references/meins.md](references/meins.md) | doc in 🔒 Meins → `Wochenreviews` | +| `ich` | [references/ich.md](references/ich.md) | doc in 🔒 Meins → `Wochenreviews`; state row in `tools/it-cockpit/state/ich/` | | `person ` | [references/person.md](references/person.md) | nothing, by hard rule | | `portfolio` | [references/portfolio.md](references/portfolio.md) | doc in 🔒 Meins → `02 Projekte & Initiativen` | @@ -94,16 +95,16 @@ Follow [references/execute.md](references/execute.md): one command per change, t ### 5. Persist and report -For `meins` and `portfolio` write the document as the mode's reference describes, reflecting what was actually executed. Other modes write nothing. +For `ich` and `portfolio` write the document as the mode's reference describes, reflecting what was actually executed. Other modes write nothing. Report in German: executed items with links, failed items with error, document URL (if any), open `Für dich:` points. -**Done when** the report is out and, for `meins`/`portfolio`, the document is readable in Basecamp. +**Done when** the report is out and, for `ich`/`portfolio`, the document is readable in Basecamp. ## Guardrails - Every write waits for the batch approval; the approval covers exactly the numbered items shown. - Proposals carry their reason; a bare "Termin verschieben" is incomplete. - OKF holds goals, context, decisions and links; task state lives in Basecamp (spec §3.2). OKF fixes go through the same approval and end with the `validate` skill on the bundle. -- Process KPIs only: counts per list or per process, never per-person rankings. +- Process KPIs only: counts per list or per process, never per-person rankings. Scores, trends and persisted metrics exist only in mode `ich` and only about Sascha; the person lens stays score-free. - Owner reminders belong to automation A3; the skill's output to colleagues is the comment on the object, nothing else. diff --git a/skills/lead-radar/references/execute.md b/skills/lead-radar/references/execute.md index 4db810d..f1fe675 100644 --- a/skills/lead-radar/references/execute.md +++ b/skills/lead-radar/references/execute.md @@ -28,6 +28,9 @@ Mention the addressees as described under **Mentions** in `SKILL.md` (`[@Name](p | Move to list (same project) | `basecamp todos position --to 1 --list --json` | | Complete | `basecamp todos complete --json` | | Read back | `basecamp todos show --json` | +| Create to-do (own commitment, mode `ich` only) | `basecamp todos create "" --in --list --assignee [--due YYYY-MM-DD] --description "" --json` | + +Creating a to-do is allowed only in mode `ich`, only for Sascha's own commitment from a Sally recording, owner always Sascha. The description carries one context sentence and the link to the recording; the new to-do gets the same why-comment as every change. **Parking:** diff --git a/skills/lead-radar/references/ich.md b/skills/lead-radar/references/ich.md new file mode 100644 index 0000000..237ae2b --- /dev/null +++ b/skills/lead-radar/references/ich.md @@ -0,0 +1,105 @@ +# Mode `ich`: weekly self-review + +When: weekly, ideally Friday after the close or Monday morning. Lens: Sascha's own commitments across **all** projects, not only 🔒 Meins. Four sections answer four weaknesses (spec `docs/2026-11-05-lead-radar-ich.md` §1): A commitments kept (Termintreue), B too much held personally, C open loops from meetings, D focus. Then a look back at the numbers. + +## Boundary (spec §2.3) + +The self-lens is the mirror image of the person lens: scores, trends and a persisted trail are allowed, because they are about Sascha only. The output never contains numbers about colleagues; delegation proposals name role and skills from the profile, not workload or comparisons. The only effect on others is the comment on a delegated to-do. + +## Input (`--scope ich`) + +```bash +python3 tools/it-cockpit/cockpit.py --json --scope ich --dry-run +``` + +Run the review on `--dry-run`; the state is written at the very end (see Document). Keys under `ich`: + +- `kennzahlen`: `offen`, `je_projekt`, `ohne_termin`, `ohne_termin_anteil`, `ueberfaellig`, `still_30d`, `verschoben` (ids whose date moved later since the last run), `mehrfach_verschoben` (`verschiebungen ≥ 3`), `erledigt`, `puenktlich`, `termintreue`, `delegiert`, `sonstig_weg`, `neu`, `vorheriger_lauf`. `vorheriger_lauf == null` is the first run: comparison fields are `null`, say "Vergleich ab dem zweiten Lauf". +- `klassifikation`: `ueberfaellig`, `heute`, `diese_woche`, `still_30d`, `ohne_termin` (Daily Quests excluded). +- `listen_je_projekt`: every open to-do, grouped project → list. +- `erledigt`: completed since the last run, each with `completed_at` (null when Basecamp gave no event). +- `verlauf`: the last four metric rows, oldest first. +- Top level: `abwesenheiten` (next 10 days), `drift`, `errors`. + +Read the OKF profiles in `/home/sascha.koenig/p/PM/.okf/organization/` before section 3, and `processes/meeting-runden.md` before section 1. + +## 1. Zusagen (C): commitments without a to-do + +1. Sally `search_appointments`: the last 7 days (back to `vorheriger_lauf` when that is longer ago), Sascha as participant. Keep recordings with `sallyStatus = transcriptionAndSummarizationCompleted`; list the others in one line as "noch nicht ausgewertet". +2. Per recording: `list_tasks` filtered to tasks assigned to Sascha, plus follow-ups from `get_summary` that name Sascha. +3. Match each commitment against `listen_je_projekt` and `erledigt` by title similarity and the round's project (round profile in `meeting-runden.md`). Match → nothing. No match → proposal **To-do anlegen**: project from the round profile (default 🔒 Meins → `02 Nächste Schritte`), owner Sascha, date from the meeting or marked `Vorschlag`, description = one context sentence + link to the Sally recording. +4. Commitments Sally assigns to other people are not this mode's business (`lead-meeting-followup`). + +## 2. Termine (A): overdue and repeatedly moved + +For every item in `klassifikation.ueberfaellig` and `kennzahlen.mehrfach_verschoben` exactly one recommendation: + +- **do**: date within this week, comment names the next step +- **defer**: a new honest date with the reason in the comment. With `verschiebungen ≥ 3` only when Sascha states a reason; otherwise propose **close** or **delegate** instead. The skill never proposes a fourth move without a reason. +- **close**: complete with a comment saying why it is no longer needed + +## 3. Abgeben (B): delegation candidates + +Candidates: open to-dos whose title or description matches role, responsibilities or skills in a profile, and the person is not absent in the next 10 days (`abwesenheiten`). Proposal: owner change + comment with context, definition of done and the mention `[@Vorname Nachname](person:)`. Topics with no matching profile: `Für dich: wem gibst du X?`. + +At most **five** delegation proposals per run; mention the rest as "weitere Kandidaten: N". + +## 4. Fokus (D): the five for this week + +Top 5 with reason (date, weight, blocks others, GF/compliance). Valid items with no room in the next two weeks → park (🔒 Meins → `05 Backlog / Später`, or the project's backlog per `execute.md`). Silent, unclear or duplicate → clarify (comment with the next step) or close. + +## Rückblick + +Table of this run's `kennzahlen` next to `verlauf` (up to four earlier runs), then two sentences: what improved, what did not. First run: current state only. + +## Template (chat and document) + +```md +# Wochenreview – YYYY-MM-DD + +## Kurzfazit +[2–4 Sätze] + +## Deine 5 für diese Woche +1. [To-do](link) · Projekt — warum jetzt / nächster Schritt + +## Zusagen ohne To-do +- Zusage · Meeting (Datum) → Vorschlag Nr. N + +## Termine +- [To-do](link) · N× verschoben — do / defer / close + +## Abgeben +- [To-do](link) → Vorname (Rolle) — Vorschlag Nr. N + +## Aufräumen +- [To-do](link) — parken / clarify / close + +## Erledigt seit letztem Review +- [To-do](link) · pünktlich / zu spät + +## Kennzahlen +| | heute | −1 | −2 | −3 | −4 | +|---|---|---|---|---|---| +| offen | | | | | | +| ohne Termin | | | | | | +| überfällig | | | | | | +| Termintreue | | | | | | +| verschoben / ≥3× | | | | | | +| delegiert | | | | | | +| neu | | | | | | + +[zwei Sätze Trend] + +## Vorgeschlagene Änderungen +1. [nummeriert wie in der Freigabeliste; Status ✅ umgesetzt / ❌ abgewählt / ⚠️ fehlgeschlagen] +``` + +## Document and state + +After execution (step 4 of `SKILL.md`): + +1. Save the review as document `YYYY-MM-DD – Wochenreview` in 🔒 Meins → folder `Wochenreviews` with the change list showing what was actually done: `basecamp files documents create "" "$(cat )" --in --vault --no-subscribe --json`. Read it back and keep its URL. A user who asks for a chat-only review gets no document. +2. Run the cockpit once more **without** `--dry-run`: `python3 tools/it-cockpit/cockpit.py --json --scope ich > /dev/null`. This writes `state/ich/snapshot.json` and the day's row in `verlauf.jsonl` with the state after execution, so next week's comparison starts from what was actually done. + +**First run:** the folder `Wochenreviews` does not exist yet (`systems/basecamp.md` says so). Add one numbered proposal "Ordner `Wochenreviews` in 🔒 Meins anlegen und ID in `systems/basecamp.md` eintragen"; after approval create it with `basecamp files folders create Wochenreviews --in --json` and replace the placeholder in the ID table with the new ID. diff --git a/skills/lead-radar/references/meins.md b/skills/lead-radar/references/meins.md deleted file mode 100644 index 40a0df1..0000000 --- a/skills/lead-radar/references/meins.md +++ /dev/null @@ -1,61 +0,0 @@ -# Mode `meins`: weekly review of Sascha's own to-dos - -When: weekly. Lens: Sascha's own to-dos in 🔒 Meins; a review with judgement and recommendations, not a list dump. - -## Input - -- `--scope meins` → `meins.listen` (open to-dos per list) and `meins.klassifikation` (`ueberfaellig`, `heute`, `diese_woche`, `still_30d`, `wartend_nachfassen`, `ohne_owner`, `ohne_termin`). -- Last review: the newest document in folder `Wochenreviews` (`basecamp files list --in --vault --json`); its date bounds "Erledigt seit letztem Review". No folder or no document yet → use the last 7 days. -- Completed to-dos: `basecamp todos list --in --completed --all --json`, keep those completed after that date. - -## Classify every open to-do (exactly one class) - -- `🔥 Diese Woche erledigen/entscheiden`: due now, strategically important, or blocking others -- `⚠️ Überfällig`: due date passed -- `⏳ Wartet auf Rückmeldung`: delegated, blocked, or needs another person (includes the Meins waiting list) -- `🧹 Aufräumen`: unclear, duplicate, stale, or no longer relevant -- `📦 Backlog`: valid but not for this week - -## Score each 🔥, ⚠️ and ⏳ item - -| Dimension | Values | -|---|---| -| Dringlichkeit | überfällig / diese Woche / später / ohne Termin | -| Klarheit | nächster Schritt klar? Definition of Done vorhanden? | -| Momentum | neu / 7+ Tage still / 30+ Tage still (`still_tage`) | -| Ownership | Sascha / wartet auf jemanden / unklar | - -The scores drive the recommendation per item: **do, defer, delegate, clarify, park, close**. Each recommendation that changes Basecamp becomes a numbered proposal (new date, move to the Meins backlog list, owner change for delegation, complete with comment, comment that rewrites an unclear item's next step). - -## Review template (chat and document) - -```md -# Wochenreview To-dos – YYYY-MM-DD - -## Kurzfazit -[2–4 Sätze] - -## Top 5 für diese Woche -1. [Todo](link) — Warum jetzt / nächster Schritt - -## Überfällig -- [Todo](link) — Empfehlung: do / defer / delegate / clarify / close - -## Wartet auf Rückmeldung -- [Todo](link) — wartet auf [Person/Sache], nächster Nachfasspunkt - -## Aufräumen -- [Todo](link) — Grund und empfohlene Aktion - -## Erledigt seit letztem Review -- [Todo](link) - -## Vorgeschlagene Änderungen -1. [nummeriert wie in der Freigabeliste; im Dokument mit Status ✅ umgesetzt / ❌ abgewählt / ⚠️ fehlgeschlagen] -``` - -## Document - -After execution (step 5 of `SKILL.md`), save the review as document `YYYY-MM-DD – Wochenreview` in 🔒 Meins → folder `Wochenreviews`, with the change list showing what was actually done. Use `basecamp files documents create "" "$(cat )" --in --vault --no-subscribe --json`. Read it back and keep its URL. A user who asks for a chat-only review gets no document. - -**First run:** the folder `Wochenreviews` does not exist yet (`systems/basecamp.md` says so). Add one numbered proposal "Ordner `Wochenreviews` in 🔒 Meins anlegen und ID in `systems/basecamp.md` eintragen"; after approval create it with `basecamp files folders create Wochenreviews --in --json` and replace the placeholder in the ID table with the new ID. diff --git a/skills/lead-radar/references/tag.md b/skills/lead-radar/references/tag.md index 3d0a005..133b4aa 100644 --- a/skills/lead-radar/references/tag.md +++ b/skills/lead-radar/references/tag.md @@ -7,6 +7,7 @@ When: in the morning after the cockpit, before the 10-minute briefing (led by Me - `zammad`: `ohne_owner`, `sla_4h`, `prio_ohne_antwort`, `absprache_ohne_datum`, `wiedervorlage_faellig` (`offen` is context only). `null` plus an `errors` entry = Zammad blind. - `basecamp_it`: `ueberfaellig`, `heute`, `diese_woche`, `wartend_nachfassen`, `ohne_owner`, `ohne_termin`, `still_30d`. - `abwesenheiten`: who is out today and in the next 10 days. +- `du_heute`: Sascha's own to-dos across all projects: `heute`, `ueberfaellig`, `mehrfach_verschoben` (date moved ≥ 3 times). `null` with `--no-xp` → write `Du heute: nicht geladen`. Read the OKF profiles of the team for role and skills before proposing an owner. @@ -22,6 +23,8 @@ Read the OKF profiles of the team for role and skills before proposing an owner. | Rule violation (no owner / no date) | set owner and date, labelled `Vorschlag` | | Wartet auf extern, follow-up date reached | new follow-up date + comment with what is being waited for | | `still_30d` | leave for Friday; mention the count only | +| Own to-do overdue (`du_heute.ueberfaellig`) | new date with the reason in the comment, or complete; a ≥3× moved item gets no fourth move without a stated reason | +| Own to-do due today | none; it shows in „Du heute“ | An absent owner is the strongest reason for an owner change; everything else first asks the owner via a comment. @@ -32,6 +35,11 @@ An absent owner is the strongest reason for an owner change; everything else fir **Kurzfazit:** <2–3 Sätze: was heute brennt, was entschieden werden muss> +## Du heute +**Fällig:** [To-do](link) · Projekt — [To-do](link) · Projekt +**Überfällig:** N, davon ≥3× verschoben: [To-do](link) +**Deine 3 Dinge heute?** Vorschlag: 1. … 2. … 3. … + ## Zammad @@ -46,3 +54,5 @@ An absent owner is the strongest reason for an owner change; everything else fir ``` Empty sections collapse to one line ("Zammad: nichts ohne Owner, kein SLA-Risiko"). + +„Du heute“ comes first: Sascha sets his own day before he looks at the team. The three things are proposed from due date and weight (blocks others, GF date, compliance); Sascha confirms or names his own, and nothing is stored. Two minutes, no extra approval round.