LinkedIn Field Note
One job: take the next AI Field Notes entry from "what should I write about?" to published on LinkedIn, reproducing the pipeline proven by hand on field note #01 (33-agent article + ultra-short post, 2026-07-21/22) and the Start-with-three toolkit (2026-07-26).
This is personal thought-leadership on Ane's personal brand. Wine & Slate tokens, warmth-inside-structure voice, self-irony allowed once per piece. It is not IPPF work: never use ane_package.reporting.brand, never the IPPF template.
The hard rule
Never publish, schedule, or stage anything Ane has not approved in the current session. Approval is per artifact (article, post, first comment, toolkit PDF, visuals) and does not carry across sessions: if publishing happens in a later session, Ane re-confirms the final file with a one-line go before anything touches LinkedIn.
Paths
| What |
Where |
| Content folder |
C:/Users/AGasser/OneDrive/2. Ane's AREAS/AG Business/LinkedIn content/ |
| Series ledger |
<content folder>/field-notes-ledger.md (create at first run: see references/ledger.md) |
| Brand tokens (canonical) |
C:/Users/AGasser/OneDrive/2. Ane's AREAS/AG Business/Brand/brand-tokens.md |
| docx/pptx helpers |
C:/Users/AGasser/OneDrive/2. Ane's AREAS/AG Business/Brand/generators/pb.py (+ gen_report.py, gen_article.py) |
| Visual spec (LinkedIn) |
2. Ane's AREAS/AG Business/Brand/BRAND-SPEC.md |
| Quality-bar exemplars |
2026-07-26 Three-Agent Starter Kit (article companion).docx and Start-with-three.docx (Ane's hand-edited final), build_three_agent_kit.py, 2026-07-22 LinkedIn POST v3 ULTRA-SHORT - 33-agent AI system.md |
Five stages
Run the stages in order for a new field note; enter mid-pipeline when Ane resumes one (the ledger records where every note stands). Read only the reference file for the stage you are in.
| Stage |
What it produces |
Read |
| 1 TOPIC |
The next topic, chosen by Ane from 2-3 ledger-grounded proposals |
references/ledger.md |
| 2 INTERVIEW |
Lessons, story, numbers, toolkit promise — extracted one question at a time |
references/interview.md |
| 3 DRAFT |
Article .md + branded .docx, short post .md, first-comment text |
references/draft-formats.md |
| 4 TOOLKIT |
The companion artifact the article promises: generator script, .docx, .pdf, cover illustration, cover visual, and the caption for the document post |
references/toolkit.md |
| 5 PUBLISH |
The posts live on LinkedIn per the agreed calendar; ledger updated |
references/publish.md |
After every stage: update the ledger entry for the note (status, paths, decisions). Milestone writes — draft and correct in conversation, write files at confirmed decision points, not per message.
Standing rules (all stages)
- Factual reliability. Never invent facts, numbers, names, or quotes about Ane, her system, or her work. Every number in a draft (agent counts, check counts, page counts, dates) is verified against source in the drafting session or flagged
⚠️ verify before posting. If a needed fact is not in loaded context, ask.
- Voice. Warmth inside structure: a two-sentence concrete story may open a piece, the finding lands by sentence three. Address the reader as "you". Plain English, translatability (a Romanian, Tunisian, Ethiopian, or Vietnamese English-speaking reader understands on first read), sentences under 25 words, active voice. No em-dashes in any prose Ane will publish. Self-irony at most once per piece, and the P.S. is usually where it lives.
- AI-use disclosure. The disclosure is mandatory on every artifact, not only articles and posts. Articles and posts carry the P.S. disclosure line; kits carry the disclosure paragraph in exemplar wording before the brand footer, docx and workbook alike; document-post captions carry the compressed P.S. The framing is always Ane's drafts-verifies-decides: written with help from the system it describes. This is the series' credibility signature, not boilerplate, and it is enforced by the Stage 4 checklist and Stage 5 pre-publish check, never by copying an exemplar generator.
- Edit preservation. Apply mel_wiki/wiki/concepts/edit-preservation-protocol.md when target file exists. Once Ane hand-edits a generated
.docx, that .docx is authoritative: subsequent changes are targeted edits (docx-revision-pass method if that skill exists, else Word COM / python-docx scoped fixes), never regeneration from the .md.
- Composition. Before showing Ane a final draft, run the ane-voice checklist over it (Tier 1 register, personal warmth preserved). Her Word revisions route through docx-revision-pass when built. The dictation intake follows the meeting-notes charitable-transcription rules; the interview follows the grill-mel recommend-then-ask pattern.
Close
End any working session on a note by telling Ane exactly where the note stands (stage, what is approved, what is pending) and what the next session starts with. The ledger carries the same state in writing.
1---2name: linkedin-field-note3description: Produce and publish the next entry in Ane's personal AI Field Notes series on LinkedIn, end to end: topic proposal from the series ledger, a dictation interview, the article plus short post, the companion toolkit, then publication via claude-in-chrome after per-artifact approval. Runs on Ane's personal Wine & Slate brand, never IPPF branding. Use for any personal AI/MEL thought-leadership piece for LinkedIn. Not for IPPF-branded MA-facing learning products (learning-product), meeting notes (meeting-notes), voice audits of existing text (ane-voice), or LinkedIn profile edits.4---5
6# LinkedIn Field Note
7
8One job: take the next AI Field Notes entry from "what should I write about?" to published on LinkedIn, reproducing the pipeline proven by hand on field note #01 (33-agent article + ultra-short post, 2026-07-21/22) and the Start-with-three toolkit (2026-07-26).
9
10This is personal thought-leadership on Ane's personal brand. Wine & Slate tokens, warmth-inside-structure voice, self-irony allowed once per piece. It is not IPPF work: never use `ane_package.reporting.brand`, never the IPPF template.
11
12## The hard rule
13
14**Never publish, schedule, or stage anything Ane has not approved in the current session.** Approval is per artifact (article, post, first comment, toolkit PDF, visuals) and does not carry across sessions: if publishing happens in a later session, Ane re-confirms the final file with a one-line go before anything touches LinkedIn.
15
16## Paths
17
18| What | Where |
19|---|---|
20| Content folder | `C:/Users/AGasser/OneDrive/2. Ane's AREAS/AG Business/LinkedIn content/` |
21| Series ledger | `<content folder>/field-notes-ledger.md` (create at first run: see `references/ledger.md`) |
22| Brand tokens (canonical) | `C:/Users/AGasser/OneDrive/2. Ane's AREAS/AG Business/Brand/brand-tokens.md` |
23| docx/pptx helpers | `C:/Users/AGasser/OneDrive/2. Ane's AREAS/AG Business/Brand/generators/pb.py` (+ `gen_report.py`, `gen_article.py`) |
24| Visual spec (LinkedIn) | `2. Ane's AREAS/AG Business/Brand/BRAND-SPEC.md` |
25| Quality-bar exemplars | `2026-07-26 Three-Agent Starter Kit (article companion).docx` and `Start-with-three.docx` (Ane's hand-edited final), `build_three_agent_kit.py`, `2026-07-22 LinkedIn POST v3 ULTRA-SHORT - 33-agent AI system.md` |
26
27## Five stages
28
29Run the stages in order for a new field note; enter mid-pipeline when Ane resumes one (the ledger records where every note stands). Read only the reference file for the stage you are in.
30
31| Stage | What it produces | Read |
32|---|---|---|
33| 1 TOPIC | The next topic, chosen by Ane from 2-3 ledger-grounded proposals | `references/ledger.md` |
34| 2 INTERVIEW | Lessons, story, numbers, toolkit promise — extracted one question at a time | `references/interview.md` |
35| 3 DRAFT | Article `.md` + branded `.docx`, short post `.md`, first-comment text | `references/draft-formats.md` |
36| 4 TOOLKIT | The companion artifact the article promises: generator script, `.docx`, `.pdf`, cover illustration, cover visual, and the caption for the document post | `references/toolkit.md` |
37| 5 PUBLISH | The posts live on LinkedIn per the agreed calendar; ledger updated | `references/publish.md` |
38
39After every stage: update the ledger entry for the note (status, paths, decisions). Milestone writes — draft and correct in conversation, write files at confirmed decision points, not per message.
40
41## Standing rules (all stages)
42
43- **Factual reliability.** Never invent facts, numbers, names, or quotes about Ane, her system, or her work. Every number in a draft (agent counts, check counts, page counts, dates) is verified against source in the drafting session or flagged `⚠️ verify before posting`. If a needed fact is not in loaded context, ask.
44- **Voice.** Warmth inside structure: a two-sentence concrete story may open a piece, the finding lands by sentence three. Address the reader as "you". Plain English, translatability (a Romanian, Tunisian, Ethiopian, or Vietnamese English-speaking reader understands on first read), sentences under 25 words, active voice. No em-dashes in any prose Ane will publish. Self-irony at most once per piece, and the P.S. is usually where it lives.
45- **AI-use disclosure.** The disclosure is mandatory on every artifact, not only articles and posts. Articles and posts carry the P.S. disclosure line; kits carry the disclosure paragraph in exemplar wording before the brand footer, docx and workbook alike; document-post captions carry the compressed P.S. The framing is always Ane's drafts-verifies-decides: written with help from the system it describes. This is the series' credibility signature, not boilerplate, and it is enforced by the Stage 4 checklist and Stage 5 pre-publish check, never by copying an exemplar generator.
46- **Edit preservation.** Apply mel_wiki/wiki/concepts/edit-preservation-protocol.md when target file exists. Once Ane hand-edits a generated `.docx`, that `.docx` is authoritative: subsequent changes are targeted edits (docx-revision-pass method if that skill exists, else Word COM / python-docx scoped fixes), never regeneration from the `.md`.
47- **Composition.** Before showing Ane a final draft, run the ane-voice checklist over it (Tier 1 register, personal warmth preserved). Her Word revisions route through docx-revision-pass when built. The dictation intake follows the meeting-notes charitable-transcription rules; the interview follows the grill-mel recommend-then-ask pattern.
48
49## Close
50
51End any working session on a note by telling Ane exactly where the note stands (stage, what is approved, what is pending) and what the next session starts with. The ledger carries the same state in writing.
gasserane