Source: frameworks/kit-tpc-post-momentum-monday/04-tpc-post-momentum-monday-quality.md

04 — QUALITY: TPC Post-Momentum-Monday Kit

When to run: After all eight outputs are drafted, before presenting to Kathryn. Format: Two gates. Gate 1 is binary (pre-build readiness). Gate 2 is per-output checklist.

Gate 1 — Pre-Build Readiness (Binary)

Cannot proceed past this gate if any check is no.

[ ] TPC reference data file (cyp/tpc/tpc-reference-data.md) has been read FIRST — proper-noun spellings, firm names, tools, prior commitments are all loaded
[ ] Transcript is readable end-to-end
[ ] All 7 member files have been read in this session
[ ] TPC CLAUDE.md has been read in this session
[ ] Facilitation guide for this session has been located and read
[ ] Planning doc current state has been read
[ ] Attendee list cross-checked against TPC roster + reference data (anyone not in transcript is flagged as absent)
[ ] Absentees' reasons identified (or marked unexplained)
[ ] Kathryn confirmed the run plan ("Confirm Understanding" step from 00-start-here)

Gate 2 — Per-Output Quality Checklists

Output 1 — Monday recap email

[ ] Subject line matches [Test] TPC: Momentum Monday Recap- M-DD-YYYY (note: [Test] prefix retained during system-in-development phase — confirm with Kathryn before removing)
[ ] Opener is Hi there, (matches MailerLite render pattern)
[ ] Recording link + chat link placeholders present in second paragraph (or actual URLs if Kathryn has them)
[ ] Brief context paragraph notes attendees + Meryl/absentee acknowledgment without listing absentees as participants
[ ] Member specifics (3-5 bullets) are celebratory/neutral content only, no exposure
[ ] In-room commitment reinforced with bolded deadline
[ ] "What I'm listening for" sub-bullets present if there's an ask
[ ] No internal facilitator language ("convergence," "pivot," "watch-fors," "pain points")
[ ] No corporate-speak from the forbidden list (file 02)
[ ] P.S. used for housekeeping (Claude subscription, scope notes) rather than a separate Housekeeping header — keeps body lighter
[ ] First-name sign-off, no "looking forward" filler
[ ] Length target hit: 350-500 words
[ ] HTML version has inline styles that render in Gmail; no external stylesheets, no scripts
[ ] HTML version uses — and & entities correctly
[ ] HTML version has placeholder anchors (...) so Kathryn can swap URLs in
[ ] Both .md and .html files exist in cyp/tpc/drafts/

Output 9 — Google Doc overview entry

[ ] Format: #### header, What We Discussed, numbered topics with sub-bullets
[ ] Starts at first substantive topic (no weekend/social check-in section)
[ ] Sub-bullet format throughout — no narrative paragraphs
[ ] No editorial commentary ("The pattern:", "The fix:", "Fair.", "Valid point.")
[ ] No thematic framing ("an accidental theme showed up," "nobody planned it but...")
[ ] Members' questions reported as questions — never framed as "pushed back," "resisted," "disagreed"
[ ] No characterization of member understanding ("where it clicked," "she tracked that but...")
[ ] Quoting proportional to weight — casual questions get paraphrases, not block-quote treatment
[ ] All member names spelled per reference data file
[ ] 5-8 numbered topics, each with 3-6 sub-bullets (not fewer, not more than 6 per topic)
[ ] Content matches transcript — no invented contributions, no extrapolation
[ ] Facts in doc overview match facts in email recap (Output 1) — no contradictions between the two

Output 2 — Personal facilitator recap

[ ] Header has Date, Meeting, Attendees (named), Absent (with reason)
[ ] Summary is 2-4 sentences capturing the session arc
[ ] "What We Discussed (in order)" section follows the actual transcript order
[ ] Each member with substantive contributions has their own subsection
[ ] Quotes are actual transcript quotes, attributed correctly (member-name spelling per terminology file)
[ ] My Takeaways section connects to Kathryn's broader business / IP / clients
[ ] Connections to Follow Up table is complete (one row per attendee + absentees)
[ ] Action Items list is concrete and dated where relevant
[ ] Ideas Sparked section names what's reusable
[ ] Quotes Worth Remembering section has speaker + context for each

Output 3 — Facilitator review

[ ] "Did the session do its job?" framing is honest, not self-protective
[ ] "Where you held the plan / where you broke from the plan" structure used
[ ] Each plan-departure has a trade-off analysis (cost / benefit / honest question)
[ ] Airtime section names dominance patterns specifically
[ ] What landed / what didn't land sections are concrete
[ ] Calibration-for-next-session has 4-6 specific actions
[ ] At least one self-question Kathryn should sit with
[ ] No sanitizing for an external audience — voice is candid

Output 4 — Member file updates

[ ] Each attendee has a new ## YYYY-MM-DD Session Contributions section appended
[ ] Section includes: AI/topic-specific contributions, pain points named, quotes for the file (if applicable), commitment made, new watch-for added
[ ] If the existing profile shifted, the existing sections are revised (not just appended)
[ ] Each absentee has a brief absence note appended with action item
[ ] No member's contributions are invented or extrapolated beyond what's in the transcript
[ ] Topic Candidates for Three Events updated where new signal warrants

Output 5 — JSON

[ ] Schema matches the 2026-04-13 baseline (meetingDate, meetingType, participants, conversation, takeaways, actions, quotes, connections, ideas)
[ ] absent array added with name + reason for any non-attending members
[ ] All names spelled per the terminology file's spelling authority
[ ] All quotes are exact transcript quotes (not paraphrased)
[ ] actions items have dates where the session set a date; empty string otherwise
[ ] connections includes both attendees and absentees with appropriate follow-ups
[ ] Valid JSON (would parse without error)

Output 6 — Planning doc append

[ ] New section appended at bottom (not interleaved into existing content)
[ ] Section title names what changed (e.g., "Arc Has Shifted" / "Arc Holds — minor adjustments")
[ ] If arc shifted: revised arc table provided with row-level changes
[ ] "What's new in the arc" and "What's still uncertain" sub-sections present
[ ] Cross-reference to facilitator review and lessons-learned (if any) included

Output 7 — Lessons-learned entry (conditional)

[ ] Decision: Was a real lesson surfaced? (If no — skip this output. Don't force one.)
[ ] If yes: file format matches cyp/tpc/lessons-learned/README.md
[ ] One-line lesson title is specific and decision-actionable
[ ] What happened section describes concretely (no generalities)
[ ] What I'd do differently section is specific (the redo)
[ ] Implication for Practice Builders mastermind section names what carries forward
[ ] Cross-reference to facilitator review at bottom

Output 8 — Friday email skeleton

[ ] Job-of-the-week named explicitly (one job, not five)
[ ] Subject line options provided (3)
[ ] Section-by-section skeleton with placeholder text
[ ] At least 2 deeper-question options provided (matching different patterns the week may produce)
[ ] "What to verify before sending" section addresses the verification step
[ ] "Do NOT" list catches the predictable failure modes
[ ] Voice matches the 4-24 Friday email pattern (direct, warm, no fluff)
[ ] Skeleton placeholder text is clearly marked as STRUCTURAL only — not voice. Add a note at top of skeleton: "The example text in [brackets] is structural reference. Do not paraphrase or polish it. When filling, voice anchors are the prior shipped Friday email and the most recent transcript — see file 05 Voice Anchor Rule."

Output 8 (filled) — When the skeleton is filled into a sent Friday email

[ ] Every line traces to either (a) the prior shipped Friday email in archive/, (b) the most recent transcript in cyp/tpc/transcripts/, or (c) Kathryn's known voice patterns from member feedback memories
[ ] No skeleton example text preserved as voice (failure mode: lifting the skeleton's [example: ...] text and treating it as Kathryn's voice)
[ ] No coach/facilitator phrases (per feedback-no-coach-language.md): "sit with," "rattle around for me," "lean in," "show up," "no prep," "scaffold," "starting place," "worth naming," "hold space"
[ ] No insider vocabulary (per feedback-tpc-no-insider-words.md): "kit" as noun, "playbook," "framework" as noun
[ ] No fabricated patterns — if listening hasn't happened, do not claim patterns from recordings
[ ] Bridge sentences explicitly connect each section to the others (Section 2 question must connect to the recordings/onboarding context, not float free)

Output 10b — Listen-pass extraction template

[ ] Recording inventory table with all 7 members listed (mark "not received" where appropriate)
[ ] Engagement read counts present (in-room vs absent-no-notice)
[ ] Per-member raw notes section for each member (at least placeholder structure for non-receivers)
[ ] Cross-cutting patterns section blank but structured (members where surfaced / how described / why matters)
[ ] Comparison-to-prior-Monday-patterns table referenced with prior patterns pre-populated
[ ] "What surprised you" + "What didn't show up" + "Confidentiality flags" sections present
[ ] Implications-for-scaffold section ties to specific decision items from action items doc
[ ] Cross-references to prior Monday personal recap, facilitator review, action items doc, Friday email sent — all linked at top

Output 11 — Next Monday's facilitation guide

[ ] Header includes date, time, expected attending, expected absent (with reason)
[ ] "The Job This Session Does" section names 2-3 things explicitly + names what is NOT being done
[ ] Pre-Session Inputs section points to the listen-pass extraction (10b), Friday email sent, scaffold/build, action items
[ ] Session flow has time allocations, member-specific watch-fors, pull-in prompts for at least each in-room attendee
[ ] Patterns named in segment 2 of the guide MATCH the patterns from the 10b extraction — not generic, not Monday-only
[ ] Capture sheet tables present and filled-in-able
[ ] "Things to Watch For" section catches the predictable failure modes for THIS session shape
[ ] "After the Call" section names within-24-hr and within-3-day next steps
[ ] "What Success Looks Like" is concrete + names the single best indicator
[ ] Calibration Pre-Notes pull from the prior facilitator review's "Calibration for next session" section

Output 10 — Kathryn's Action Items

[ ] Markdown action-items doc exists at cyp/tpc/sessions/tpc-action-items-YYYY-MM-DD.md
[ ] Doc has three sections: Build/listen, Decision, Doc-update
[ ] No Member-touch items (personal reach-outs, 1:1 emails) appear anywhere in the output — these are explicitly out of scope per the cadence rule
[ ] No Gmail drafts created outside the Monday/Friday cadence — only Outputs 1 and 8 produce member-facing emails
[ ] Each Build/listen item has a proposed calendar event with title, description, duration, target date
[ ] Calendar events were created via the calendar MCP (or proposed for Kathryn to create if MCP unavailable)
[ ] Each Decision item has a by-when date and a "what's needed to decide" line
[ ] Doc-update items list what's already done by the kit AND what (if anything) requires Kathryn's manual touch (e.g., reference data updates)
[ ] Items extracted from Output 2 (personal recap), Output 3 (facilitator review), Output 6 (planning doc) — sources cross-checked
[ ] Items not fitting any category are flagged for Kathryn's review, not silently dropped

Cross-Output Consistency Checks

[ ] Member names spelled identically across all outputs (cross-reference terminology file)
[ ] The Monday recap email's pattern reflection matches what the personal recap captured (no drift)
[ ] The facilitator review's deviations match what the lessons-learned entry (if any) addresses
[ ] The planning doc update's arc-shift narrative matches the facilitator review's deviation analysis
[ ] The Friday email skeleton's deeper question carries an actual missing thread from this session (not generic)

Common Failure Modes

Failure	What Happens	How to Fix
Cadence rule violated — kit drafts a member-facing email outside Monday/Friday	Member experience inconsistent; trust erodes	Hard-constraint check on every Output 10 run: zero member-facing emails permitted outside Outputs 1 and 8
Reference data not read first — names misspelled, tool references wrong	Recap quality drops; trust erodes	Gate 1 enforces reference-data-first; if Gate 1 fails, run cannot proceed
Member quoted in Monday recap email	Member feels singled out	Move the quote to personal recap; pattern-level only in member-facing
Forbidden term ("convergence," "pivot," "watch-fors") leaks into Monday email	Sounds clinical	Run a forbidden-term scan against the email before HTML conversion
Member-file profile not revised when it should be	Stale member context affects future facilitation	Compare current contributions to existing engagement profile — if they conflict, revise existing sections, not just append
Lessons-learned entry forced when no real lesson surfaced	Folder bloats with non-lessons	Decision-gate at top of output 7: was a real lesson surfaced? Skip if no.
Planning doc append duplicates info from member files	Maintenance burden, two sources of truth	Member-specific stays in member files; planning doc is arc-level only
Friday email skeleton's deeper question is generic	Wastes the touchpoint	Tie the deeper question to a specific thread the week missed (e.g., avoided-conversation when Round 2 was skipped)
Skeleton placeholder text lifted as voice when filling Output 8	Email reads coach-y / not-Kathryn; multi-round rewrite churn	Voice anchor rule (file 05): voice traces to prior shipped piece + transcript ONLY. Skeleton bracketed example text is structural marker, never voice reference. If draft uses any phrase from the bracketed examples, throw it out and rewrite from anchors.
Output 8 sections disconnected from each other	Question floats free of recordings/onboarding context; reads as random reflective question	Bridge sentence explicitly connects Section 2 (question) to Section 1 (recordings) — e.g., "You sent me the operational side. Every onboarding I've built has had a conversation underneath it that hadn't happened yet. So — [question]."
Output 11 produced before listen-pass populated	Facilitation guide's pattern reflection is generic / Monday-only / fabricated	Output 11 produced AFTER 10b has content. Kit produces a structural scaffold the day of the prior Monday with placeholders; Sunday-before-next-Monday is when 11 fills with actual patterns.
Output 11 doesn't pull from 10b	Next Monday's session flow doesn't reflect what came in over the week	Required input check at top of Output 11 production: 10b file must exist AND have at least 1 cross-cutting pattern populated. If not, escalate to Kathryn before producing.
Weekend/personal check-in included in Output 9	Google Doc overview is member-facing — personal updates make it read like a journal, not a session record	Output 9 rule #8: start at first substantive topic. Social/personal opening is NOT session content. Weekend details belong in Output 2 (personal recap) only.
Facilitator observations or internal language in Output 9	Doc overview reads like facilitator notes — room-dynamic observations ("Room initially quiet"), assignment narration ("Assigned Linda to book a one-on-one"), methodology language ("process pattern," "extract"), meta-observations about how the room responded	Rules #10 and #11. Remove all facilitator-perspective content: room dynamics, assignment narration, internal vocabulary, topic titles that frame a member's problem with internal language. These belong in Output 2 (personal recap), not the group-facing record.
Output 9 written as meeting minutes instead of authored record	Every bullet starts with "Member: [thing]" — reads like a transcript index, not a curated record of what was discussed. Kathryn attributed as "Kathryn:" like any other participant. No synthesis, no variation in structure across topics.	Kathryn is the author, never attributed as "Kathryn:" — her contributions are the narrative voice. Mix attributed member quotes with unattributed synthesis. Combine related contributions into single bullets. Vary structure across topics (some quote-heavy, some synthesis, some mixed). Compare to the 5-4 doc overview as the format reference.
HTML email has scripts or external stylesheets	Doesn't paste cleanly into Gmail	Inline styles only; no scripts
Absentee not noted in Monday email	Someone wonders if they were forgotten	Always include a housekeeping line for absentees, even just "if you'd like to send a recording, same deadline"
Action items output includes member-touch items	Cadence rule violated	Hard scope: only Build/listen, Decision, Doc-update categories permitted in Output 10
Calendar event proposed without realistic duration	Kathryn ignores it	Each event has a duration tied to the actual work (e.g., listen-pass = 1 hour for 4-7 recordings; kit-build = 4-6 hour weekend block)

Pass Criteria

All eight outputs pass their per-output checklist + all cross-output consistency checks pass + Common Failure Modes scan returns clean.

If conditional outputs (lessons-learned) skipped: log the skip with reason in the Mode 2 improvement note for kit-builder.