← Vault Index
frameworks / kit-tpc-post-momentum-monday / 05-tpc-post-momentum-monday-output-skill.md
Source: frameworks/kit-tpc-post-momentum-monday/05-tpc-post-momentum-monday-output-skill.md

05 — OUTPUT SKILL: TPC Post-Momentum-Monday Kit

Scope

Produces: The full post-session artifact set from a single TPC Momentum Monday transcript (see file 00). Audience: Mixed — TPC members for outputs 1 and 8; Kathryn for outputs 2-7. Filename pattern: Per-output, see file inventory in 00. Lifecycle: One run per Momentum Monday.

Required Inputs (restated for standalone readability)

  1. TPC reference data file — cyp/tpc/tpc-reference-data.md. Read FIRST, before the transcript. Canonical authority on proper-noun spellings, firm names, tools, and prior commitments. Every name and tool reference in any output must match this file.
  2. Monday transcript (.txt or .vtt) — ground truth
  3. TPC CLAUDE.md — cohort context
  4. All 7 member files
  5. Session prep doc, if one exists (cyp/tpc/drafts/tpc-monday-prep-*.md or an events/ prep file) — may be absent; note it and proceed
  6. Planning doc current state
  7. (Recommended) Prior Friday email + prior Monday recap for voice reference
  8. The session's Zoom chat — saved locally in the meeting folder (Documents/Zoom/[date] .../meetingsavednew_chat.txt). Pull it and copy it into transcripts/ next to the transcript. It is the canonical record of anything Kathryn pasted in-session — skill download links, resources — which the audio transcript does NOT contain, and it feeds the Output 9 Chat section. (Added 2026-08-03: skill links live in the chat, not the transcript.)

Workflow Overview

This kit runs as a sequence of dependent steps. Per-output detail lives in file 06 (instructions). This file gives the orchestration.

Phase 1 — Read (no writing yet)
    1.1  Read all 7 member files
    1.2  Read TPC CLAUDE.md
    1.3  Read the session prep doc if one exists (may be absent — note and proceed)
    1.4  Read planning doc current state
    1.5  Read transcript end-to-end
    1.6  Cross-check transcript speakers against TPC roster — identify attendees + absentees
    1.7  Run Gate 1 (file 04) — confirm pre-build readiness
    1.8  Confirm with Kathryn before proceeding

Phase 2 — Extract
    2.1  Member-by-member contribution extraction (per-attendee)
    2.2  Quote harvest (with speaker + context)
    2.3  Action items (Kathryn's, members', cohort-level)
    2.4  Plan-vs-actual analysis (transcript vs. the session prep doc, if one exists)
    2.5  Lessons-learned trigger check (was a real lesson surfaced?)

Phase 3 — Produce (in this order)
    3.1  Output 9 — Google Doc overview entry (depth source — produce FIRST so 1 can summarize from it)
    3.2  Output 1 — Monday recap email (member-facing — most time-sensitive)
    3.3  Output 2 — Personal facilitator recap
    3.4  Output 3 — Facilitator review
    3.5  Output 4 — Member file updates (attendees first, then absentee notes)
    3.6  Output 5 — JSON
    3.7  Output 6 — Planning doc append
    3.8  Output 7 — Lessons-learned entry (skip if not warranted)
    3.9  Output 8 — Friday email skeleton
    3.10 Output 10 — Kathryn's Action Items (after all the recap/review work — extracts items from the prior outputs)
    3.11 Output 11 — Next Monday's prep doc (a SHORT one-pager, same day)

Phase 4 — QC
    4.1  Run Gate 2 (file 04) per-output checklists
    4.2  Run cross-output consistency checks
    4.3  Run Common Failure Modes scan

Phase 5 — Present
    5.1  Surface all artifacts to Kathryn with computer:// links
    5.2  Note any conditional outputs skipped (e.g., lessons-learned not warranted)
    5.3  Flag any gaps or judgment calls made

Content Rules (apply across all outputs)

  1. Read the transcript. Not a recap. Not a summary. Same rule as kit-post-session-production. Every factual claim traces to the transcript.
  2. Member files are the spelling authority. Cross-reference every name. Transcripts misspell.
  3. Filtering by audience is non-negotiable. Member-facing outputs use member-friendly framing. Facilitator-internal outputs are candid.
  4. Forbidden terms (file 02) never appear in member-facing outputs. Run a scan before finalizing.
  5. Don't invent contributions. If a member didn't speak on a topic, don't extrapolate from prior context.
  6. Pattern-level in member-facing, member-level in facilitator-internal. The Monday recap reflects patterns ("the room landed on..."), the personal recap names members specifically.
  7. Absentees are noted respectfully. Always include a housekeeping line in the Monday email; member-file gets a brief absence note with action item.
  8. Skip the lessons-learned output if no real lesson surfaced. Forced lessons bloat the folder.
  9. The session prep doc is the comparison source for the facilitator review, not the production source. The prep doc named a job; the transcript shows what actually happened. The review explores the gap. If no prep doc exists, say so in the review and assess against what the session was evidently for — do not invent a plan to measure against.
  1. Members own their own setup — never propose that Kathryn chase it. Their IT, machines, admin permissions, connector state, and scheduled runs are theirs. The cascade records what happened in the member file and on the board; it does NOT generate action items or calendar blocks for Kathryn to fix a member's environment. If they want help, they ask. (Set by Kathryn 2026-07-21.)

Voice / Tone Per Output

OutputVoiceAnchor Reference
1. Monday recap emailMember-facing warm — direct, collegial, understated, no fluffcyp/tpc/drafts/tpc-monday-recap-2026-06-08.md (as-sent golden — warm opener, ONE framing highlight, absentees combined into one forward-looking line, no telegraphing, eased superlatives)
2. Personal facilitator recapFacilitator-internal candidcyp/tpc/group/tpc-momentum-monday-2026-04-13.md (baseline) + 2026-04-27.md (extended)
3. Facilitator reviewYour-eyes-only honestcyp/tpc/group/tpc-momentum-monday-2026-04-27-facilitator-review.md
4. Member file updatesStructural / factual appendcyp/tpc/members/linda-stapf.md (4-27 update example)
5. JSONStructuralcyp/tpc/group/tpc-momentum-monday-2026-04-27.json
6. Planning doc appendFacilitator-internal candidcyp/tpc/planning/tpc-communication-arc-through-mastermind.md (post-4-27 update)
7. Lessons-learnedReflective, decision-actionablecyp/tpc/lessons-learned/lesson-listening-to-building-pivot-2026-04-27.md
8. Friday email skeletonPer planning/tpc-friday-email-rules.mdlatest shipped TPC: Weekly Review and Preview (MailerLite / archive/)
11. Next Monday's prep docFacilitator-internal candid (your-eyes-only)cyp/tpc/events/2026-07-in-person/mm-2026-07-20-operating-week.md (the short shape) + cyp/tpc/drafts/tpc-monday-prep-2026-07-27.md

Voice Anchor Rule (applies to ALL member-facing outputs)

Voice always traces to the prior shipped piece OR the transcript. The skeleton's placeholder example text is structural reference only — never lifted as voice.

Specifically: Output 8 (Friday email skeleton) is governed by cyp/tpc/planning/tpc-friday-email-rules.md and drafts the personal message section directly from voice anchors — it does not emit [OPENER — example: ...] placeholder text. If any skeleton output ever does include bracketed example text, treat it as a structural marker only, never a voice reference — AI-generated examples drift toward coach/facilitator phrasings that don't match Kathryn's actual voice.

When filling any skeleton (Output 8 → sent Friday email, or any other skeleton output), voice anchors are:

  1. The most recent shipped piece of the same type (e.g., the prior Friday email in archive/)
  2. The most recent relevant transcript (cyp/tpc/transcripts/)
  3. Member feedback memories (feedback-no-coach-language.md, feedback-tpc-no-insider-words.md, feedback-no-honest-word.md)

Do not paraphrase or polish the skeleton's example text. Throw it out. Build from voice anchors.

Per-Output Section Specifications

These are summary specs. Detailed step-by-step procedures live in file 06.

Output 1 — Monday recap email

Recalibration (baked from the 6-8-2026 sent version — Kathryn's edits to the kit's first draft):

  1. Warm opener beat. A short human line before the recap (e.g., "Welcome to June!"), then "Here's your recap…"
  2. Highlights = ONE framing bullet, not a tactical list. Say what the session was at a frame level. Do NOT list teaching mechanics (file formats, where-skills-go, step sequences) — those belong in co-working / the walk-through, never the member email.
  3. Absentees get ONE combined, forward-looking line — not one line each, and not "go watch the recording." Promise the on-ramp ("I'll put together a walk-through for you to use when you're ready") + offer co-working / 1:1 help.
  4. Don't telegraph going deeper in the next session. Frame co-working as "bring what you're stuck on," never "send me your transcript and run the whole thing again."
  5. Ease the superlatives and quotes. Member-facing copy is understated ("in seconds," not "two seconds"; skip the hype quotes). The proof-point enthusiasm lives in the internal recap (Output 2), not here.
  6. Don't put Kathryn's words in their mouths, and don't make it about her. Report what members did in plain terms; don't attribute her interpretations to them or center her narrative.

Locked structure (per 6-8-2026 sent recap — the new golden):

Hi there,

[Optional warm opener — one short line, e.g. "Welcome to June!"]

Here's your recap from this week's Momentum Monday.

If you want to hear the conversation, here's the link. If you'd like to see the session notes, here's the link.

**What we covered:** [ONE–2 sentences — the session arc]

**Highlights:**
- [ONE framing bullet — what the session was, at altitude. Not a list of mechanics.]

**What Surfaced**
- Member: [ONE line of substance — present members, names + a sentence. Personal touches OK; acknowledge a significant life moment briefly.]
- [Absent members, combined]: [ONE forward-looking line — walk-through coming + offer of co-working/1:1 help. Not "catch up on the recording."]

**Housekeeping:** [Co-working framed as "bring what you're stuck on" + Calendly link. Kept tight. Don't promise depth.]

Talk soon,
Kathryn

Hard rules:

Files produced: .md source + .html in cyp/tpc/drafts/ (engine copies, filename tpc-monday-recap-YYYY-MM-DD.{md,html}). Deliverable (2026-08-03 change): a readable .docx to the TPC Drive — G:/Shared drives/CREATING YOUR PLAN/TPC/Internal — Kathryn only/Working/TPC Momentum Monday Recap — Week of M-D-YYYY.docx — with the body as clean formatted text and real hyperlinks. Kathryn pastes into MailerLite from the .docx. This supersedes the HTML-only delivery (MS-write block lifted; see planning/tpc-friday-email-rules.md rule 11). Once sent, the .docx moves Working → _Archive.

Output 2 — Personal facilitator recap

Output 3 — Facilitator review

Output 4 — Member file updates

Conditional (since 2026-06-22): The TPC Member Pipeline board (Output 12) now carries each member's per-session signal (Last touch + Signal/why), so the full member-file append runs ONLY for members who moved this week — a new build need, a re-engagement, a lane/read shift, or a substantive new signal. Members merely present with nothing new are carried by the board; skip their append. (Board = the weekly per-member touch; the file = the deep record, updated on change.)

Output 5 — JSON

Output 6 — Planning doc append

Output 7 — Lessons-learned entry (conditional)

Output 8 — Friday email skeleton

Governed by cyp/tpc/planning/tpc-friday-email-rules.md (source of truth — the rules file wins on any conflict).

Output 9 — Google Doc overview entry

Do NOT (Output 9 specific — these are the recurring failure modes):

  1. Don't editorialize or summarize patterns. No "The pattern:" wrap-ups at the end of sections. No "The fix:" prescriptions. No thematic conclusions. Report what happened, stop.
  2. Don't frame questions as pushback. If a member asks a clarifying question or says they're confused, report what they said — don't characterize it as "pushed back," "resisted," or "crossed her arms." Questions are questions, not opposition.
  3. Don't characterize member understanding. Never write "where it clicked and where it didn't" or "she tracked that but..." — this implies a judgment about their comprehension. Just report what they said.
  4. Don't over-quote casual conversation. Match the weight of the quoting to the weight of the moment. A casual question gets a paraphrase or a short inline quote, not a block-quote treatment that makes it read like testimony.
  5. Don't add editorial assessments. No "valid design requirement," no "fair," no "good point." The doc is a record, not a commentary track.
  6. Don't write narrative paragraphs. Every numbered topic uses sub-bullets. Never a wall of prose. The 4-27 golden example format is sub-bullets — not paragraphs with inline quotes strung together.
  7. Don't add thematic frames. No "an accidental theme showed up," no "nobody planned it but..." The reader decides whether there's a theme. You report what was said.
  8. Don't include the weekend check-in. Social/personal opening conversation is not session content. Start with the first substantive topic.
  9. Don't editorialize section titles. Titles describe the topic, not the significance. "Onboarding vs. Pre-Onboarding" — not "Tracy's Question That Locked the Vocabulary." "Five Patterns from the Recordings" — not "Five Patterns — Walked Through with the Group." No narrative flourish in headings.
  10. Don't narrate facilitation moves. No "I turned it to the room," "I asked them to recap instead of doing it myself," "I walked through the example." No room-dynamic observations ("Room initially quiet," "the energy shifted," "the room landed on"). No assignment narration ("Assigned Linda to book a one-on-one," "tasked Bev with reviewing"). This is a record of what was discussed, not a play-by-play of Kathryn's methodology or a facilitator's observation log. If a member said something, attribute it to them. Don't explain why Kathryn prompted it, how the room reacted, or what Kathryn assigned afterward.
  11. Don't use internal language in the member-facing doc. "The trigger," "confirmed the trigger," "raised it as its own gap," "extraction conversations," "process pattern," "extract" — members don't think in these terms. Use plain language they'd use themselves. Also applies to topic titles — "Linda's Prospect-to-Client Gap" frames a member's problem with internal vocabulary. "Paid Discovery and the Handoff" describes the same topic in words they'd use.
  12. Don't orphan a quote as a closing line. If a quote doesn't have a speaker and context anchoring it to the discussion, cut it. A standalone inspirational-sounding quote at the end of a section is editorial decoration.
  13. Don't dwell on individual stumbles or "snags." If something needed fixing or setting up (e.g., outputs needed to come out as documents, not .md), frame it as "here's what we set up / made native for you" — a feature of the setup, not a member's problem or a play-by-play of who hit what. Don't over-rotate on one person's experience.
  14. Keep it thin. This is a record, not a transcript. If you're past ~4 topics or restating the whole session, cut. "Depth" for a member who wasn't there means clarity, not completeness. Don't put Kathryn's words in members' mouths, and don't center the facilitator.
  15. Link canonical docs; don't restate them. When a topic is fully covered by an existing doc (the retreat agenda, a shared guide), the entry LINKS it with 1–2 headline bullets — it never re-lists that doc's contents. Duplicating a linked doc's detail is the #1 way this entry bloats. (7-6-2026: Kathryn cut the whole retreat prep + agenda section down to a link to the agenda doc + two bullets.)

Output 10 — Kathryn's Action Items

  1. A markdown action-items doc at cyp/tpc/group/tpc-action-items-[YYYY-MM-DD].md — three sections (Build/listen, Decision, Doc-update), each item with: what it is, by-when, status (drafted / scheduled / pending decision / kit-handled).
  2. Gmail drafts when warranted — but ONLY for actions that fit the 2-email-per-week cadence. The Monday recap (Output 1) and Friday email (Output 8) already exist as Gmail drafts; the action items output should never produce additional member-facing emails.
  3. Calendar events proposed for deep-work blocks — build time, research/answer-finding, decision-by deadlines. Use the calendar MCP to create them. Each event has: title, description (what's being done, what inputs needed, what the output is), duration, target date.
  1. Read Output 2 (personal recap) — the "Action Items" section there is the source list
  2. Read Output 3 (facilitator review) — the "Calibration for next session" section adds items
  3. Read Output 6 (planning doc append) — the "what's still uncertain" section may surface decisions
  4. For each item, classify into one of the three categories. If it doesn't fit a category, it's not a kit-produced action item — flag it for Kathryn's review.
  5. For each Build/listen item, propose a calendar event with realistic duration and target date.
  6. For each Decision item, write the by-when and what's needed to decide.

Output 11 — Next Monday's prep doc

  1. Header — date, time, facilitator, one line on what this session is
  2. Why this session — 2–3 sentences. What the job is, and explicitly what it is NOT.
  3. Run it — the segments, with rough minutes. Short bullets.
  4. Watch for — per-member notes that actually matter this week. Not every member every week.
  5. Decisions to land — pointers to the action-items doc.

Output 12 — Notion board writes (final step, Review-first)

  1. KHB Operations open-items card (data source bb926f81-8fc4-4a5b-b135-3f4ebcbe93ce, tool notion-create-pages): ONE action entry per run — Title TPC [M/D] Momentum Monday — open items, Status In progress, Owner Jules, Ball With Kathryn, Workstream TPC, Type Task. Body = ## Open items checklist (every still-open item, each linked to its source) → ## Artifacts → ## Feedback.
  2. TPC Member Pipeline per-member update (data source 58a7957a-2cfd-4c0c-b68b-b30d5b616c02, tool notion-tpc-board API-patch-page; private board): for each member present (+ a light note for absentees), patch Last touch and append to Signal / why.

The Email + Doc Relationship

Output 1 (email) and Output 9 (doc overview) are paired. The email is the teaser; the doc is the substance. Every email has the same two consistent links:

When the kit's recording-link and doc-link are stable, the email can have these URLs hard-coded rather than as [RECORDINGLINK] / [GOOGLEDOC_LINK] placeholders. Until Kathryn provides the actual URLs, use placeholders.

Mode 2 update (when URLs are confirmed): Hard-code the URLs in file 06 (instructions) so future runs auto-populate them.

Delivery Checklist (the final gate before sharing with Kathryn)

Don't Interrupt the Run

When testing this kit on a new session:

  1. Run every output without mid-process correction
  2. Compare to golden example (file 03)
  3. Document gaps between actual and expected
  4. Fix the kit files via Mode 2 — not just the run's output
  5. Re-run on the next Monday to verify the fix held

This is from kit-builder methodology. Mid-process corrections mask gaps; let the kit fail completely so the failure is real.

What This Kit Does NOT Cover

Note: Calendar event creation IS now covered for Output 10's Build/listen deep-work blocks. The kit creates the events; Kathryn confirms or adjusts.

Mode 2 Wrap-Up

Step 0 — learn from what was actually SENT (do this FIRST, at the start of the next run)

Kathryn edits almost every email before sending — the draft is not the final. Don't ask whether she changed anything; go read what actually went out in MailerLite and diff it.

  1. Find the last sent TPC campaign via the MailerLite tools — list_campaigns(status: "sent"), match by name + date (e.g., "TPC: Momentum Monday Recap - " or "TPC: Weekly Review and Preview - ").
  2. Pull its content with getcampaign — diff the HTML body against the draft/skeleton this kit produced. (Note: the campaign's plaintext field is only MailerLite's "can't display HTML" fallback — use the HTML body or the screenshoturl, not plaintext.)
  3. Ignore wording/synonym tweaks (noise). Note any structural change — a section removed, an obligation softened to an offer, jargon stripped, the close changed.
  4. If a structural change looks like a repeatable pattern (not a one-off taste tweak), run Mode 2 — update the golden-example pointer to the sent version + the relevant output-skill instruction. If unsure it will recur, don't harden it into a rule — flag it for confirmation next run.
  5. If no prior sent campaign exists, or nothing structural changed, move on.

This replaces the old "did Kathryn change anything by hand?" self-report (item 1 below): Mode 2 now reads what landed in MailerLite instead of asking.

Then, after the run, also ask:

  1. ~~Did Kathryn change any output by hand?~~ → handled by Step 0 above (reads the sent MailerLite campaign).
  2. Did QC miss something Kathryn caught? → Mode 2 to update file 04
  3. Did the kit produce something it shouldn't have, or fail to produce something it should? → Mode 2 to update this output skill or add to file 06

If all are no, the run was clean. Move on.