sync-v2-application

Metadata

Status	abandoned ‖ paused
Agent identity	`3184716484e6f0ea08bb13539daf07686ee79d440505f1fdf2de0357707034c3`
Created	2026-05-01T19:42:15.045245880+00:00
Tags	`grant,urgent,sync,v2-pivot`, `eval-scheduled`

Description

This task is PAUSED on creation. It runs only after Erik has read the synthesis (~/poietic.life/notes/deep-synthesis-why-this-science-20260501.md) and explicitly resumes it.

The synthesis task identifies, in its Section F, specific places in workgraph_google_application_FINAL_v2.md where the synthesis would sharpen existing prose. This task makes those edits — surgically, without rewriting. The audit-fix task ('apply-submission-audit') runs first to avoid edit conflicts on the same file.

This is NOT a rewrite. It is a sharpening. Erik is hours from pasting into the form. Anything that destabilizes word counts or overall flow is a regression.

What to do

Read ~/poietic.life/notes/deep-synthesis-why-this-science-20260501.md (produced by deep-synthesis-why task).
Read workgraph_google_application_FINAL_v2.md AS IT NOW STANDS (after audit fixes have landed).
Read CLAUDE.md 'Word Limits' subsection so caps are exact.
For each application section the synthesis Section F flagged:
- Read the current text in that section.
- Identify the minimal edit that incorporates the synthesis sharpening.
- Make the edit.
- Recount words and confirm the section is still under its cap.
If the synthesis Section F said 'v2 already lands this synthesis, no sync needed', do nothing to the application file. Just wg log confirming that.

Constraints

Surgical edits only. If a section needs more than ~2 sentences changed, STOP and wg log that the change is too large for a same-day sync; flag it for post-submission revision instead.
Word caps are hard. After every edit, re-count.
Style: no em-dashes; no 'PI' language; no v1/KRAS terms; no 'wrote this proposal with WorkGraph' recursion claim; CRISPRme/Casgevy framing per CLAUDE.md (independent identification, not direct collaboration).
Founder order Erik / Luca / Vaughn preserved wherever the team is mentioned.
Commit the change on the agent's worktree branch with a clear message. Do not push to a remote.

Output

Write ~/poietic.life/notes/sync-application-to-synthesis-20260501.md summarizing:

Which application sections were touched
Before/after diffs (concise) for each
Word counts before and after for each touched section
Anything the synthesis flagged that was too big for a surgical edit and was deferred

wg log a one-paragraph summary on this task.

Validation

Synthesis doc actually read
Application file read in its current (post-audit-fix) state
Every touched section is still within its CLAUDE.md word cap
No em-dashes introduced
No v1 / KRAS / PI language introduced
Founder order preserved
Notes doc at ~/poietic.life/notes/sync-application-to-synthesis-YYYYMMDD.md
Changes committed on worktree branch (not pushed)

## Description

This task is PAUSED on creation. It runs only after Erik has read the synthesis (`~/poietic.life/notes/deep-synthesis-why-this-science-20260501.md`) and explicitly resumes it.

The synthesis task identifies, in its Section F, specific places in `workgraph_google_application_FINAL_v2.md` where the synthesis would sharpen existing prose. This task makes those edits — surgically, without rewriting. The audit-fix task ('apply-submission-audit') runs first to avoid edit conflicts on the same file.

This is NOT a rewrite. It is a sharpening. Erik is hours from pasting into the form. Anything that destabilizes word counts or overall flow is a regression.

## What to do

1. Read `~/poietic.life/notes/deep-synthesis-why-this-science-20260501.md` (produced by deep-synthesis-why task).
2. Read `workgraph_google_application_FINAL_v2.md` AS IT NOW STANDS (after audit fixes have landed).
3. Read CLAUDE.md 'Word Limits' subsection so caps are exact.
4. For each application section the synthesis Section F flagged:
- Read the current text in that section.
- Identify the minimal edit that incorporates the synthesis sharpening.
- Make the edit.
- Recount words and confirm the section is still under its cap.
5. If the synthesis Section F said 'v2 already lands this synthesis, no sync needed', do nothing to the application file. Just `wg log` confirming that.

## Constraints

- Surgical edits only. If a section needs more than ~2 sentences changed, STOP and `wg log` that the change is too large for a same-day sync; flag it for post-submission revision instead.
- Word caps are hard. After every edit, re-count.
- Style: no em-dashes; no 'PI' language; no v1/KRAS terms; no 'wrote this proposal with WorkGraph' recursion claim; CRISPRme/Casgevy framing per CLAUDE.md (independent identification, not direct collaboration).
- Founder order Erik / Luca / Vaughn preserved wherever the team is mentioned.
- Commit the change on the agent's worktree branch with a clear message. Do not push to a remote.

## Output

Write `~/poietic.life/notes/sync-application-to-synthesis-20260501.md` summarizing:
- Which application sections were touched
- Before/after diffs (concise) for each
- Word counts before and after for each touched section
- Anything the synthesis flagged that was too big for a surgical edit and was deferred

`wg log` a one-paragraph summary on this task.

## Validation
- [ ] Synthesis doc actually read
- [ ] Application file read in its current (post-audit-fix) state
- [ ] Every touched section is still within its CLAUDE.md word cap
- [ ] No em-dashes introduced
- [ ] No v1 / KRAS / PI language introduced
- [ ] Founder order preserved
- [ ] Notes doc at `~/poietic.life/notes/sync-application-to-synthesis-YYYYMMDD.md`
- [ ] Changes committed on worktree branch (not pushed)

`sync-v2-application`

Metadata

Description

Description

What to do

Constraints

Output

Validation

Depends on

Required by

Log