Skip to content

docs: external review guidelines + hardening self-review prompt - #99

Merged
cursor[bot] merged 5 commits into
devfrom
pr/external-review-harden-pass-a6c7
Aug 13, 2026
Merged

docs: external review guidelines + hardening self-review prompt#99
cursor[bot] merged 5 commits into
devfrom
pr/external-review-harden-pass-a6c7

Conversation

@sdntsng

@sdntsng sdntsng commented Aug 11, 2026

Copy link
Copy Markdown
Contributor

Summary

After the V3 hardening cutover (dev 1f2f26f, production main 308d3a1), the external-review handoff surfaces were stale (still described #93#97 as open / “never merge to main”). This PR revises the intake SOP, assessment pack, program status, and qualification notes, and adds a dedicated copy/paste self-review prompt for the next external reviewer.

Also completes the post-cutover staging full qualification pass pins.

Changes

  • docs/research/external-reviews/2026-08-11-hardening-self-review-prompt.md — self-contained reviewer prompt (pins, claims discipline, required reading, mandatory return format)
  • docs/research/external-reviews/README.md — current priority after hardening cutover + how to commission the next pass
  • docs/research/external-reviews/TEMPLATE.md — prompt-question answers + claims checklist
  • docs/research/external-agent-assessment-pack.md — point to new prompt; update measurements/implication/priorities
  • docs/research/program-execution-status.md — mark fix: serialize extraction claims per (tenant, subject) #93chore: V3 hardening integrate + Gate 0 staging baseline #98 merged; tips 308d3a1 / 1f2f26f
  • docs/research/README.md — headline + handoff pointer
  • Post-cutover staging pins on deploy 1f2f26f:
    • OpMem 13/13
    • Marketing passed
    • LoCoMo 1×30 15/30 (MH 50%, OD 25%) — dip vs Gate 0 18/30
    • LoCoMo 3×90 33/90 (MH 22.2%, OD 42.9%) — roughly flat vs Gate 0 32/90
  • LME abort honesty note (path proven; not publishable)

Test plan

  • Staging Render live on 1f2f26f (API + worker)
  • OpMem staging post-cutover 13/13
  • Marketing staging post-cutover passed
  • LoCoMo staging 1×30 re-pin 15/30
  • LoCoMo staging 3×90 re-pin 33/90
  • Spot-check self-review prompt links resolve in-repo
  • Claims discipline forbids inventing LME accuracy / calling post-cutover 1×30 an improvement vs Gate 0

Linear / tracking

Hardening follow-through / external re-review handoff after #93#98 merge (no new Linear ID in this docs pass).

Risk / follow-ups

  • LME-20 --publish --product-recall aborted on extraction job failure — still not publishable; needs clean isolated retry.
  • 1×30 dipped vs Gate 0 while 3×90 stayed roughly flat — hop-join tradeoff should be adjudicated in the external self-review.
  • Do not blend Gate 0 (9bad898) pins with harden-local or post-cutover pins.
  • Production already cut over at main 308d3a1 (explicit prior ask). This PR is docs/pins for dev merge.
Open in Web Open in Cursor 

sdntsng and others added 5 commits August 10, 2026 11:59
Co-authored-by: Siddhant Singh <sdntsng@users.noreply.github.com>
Land subject-ordered claims, authoritative ops, LME product-recall
harness, hop executor V2, truthful hybrid status, and Gate 0 /
qualification docs from staging (dev@1f2f26f).

Co-authored-by: Siddhant Singh <sdntsng@users.noreply.github.com>
Point intake SOP and assessment pack at the post-cutover state (hardening
merged to dev/main), refresh claims discipline, and add a dedicated
copy/paste self-review prompt for the external reviewer. Record post-cutover
OpMem/marketing staging pins and LME abort honesty.

Co-authored-by: Siddhant Singh <sdntsng@users.noreply.github.com>
Record staging deploy 1f2f26f LoCoMo 1x30 at 15/30 with honest compare
vs Gate 0 18/30 and harden-local 14/30; refresh handoff surfaces.

Co-authored-by: Siddhant Singh <sdntsng@users.noreply.github.com>
Complete full-pass pins on deploy 1f2f26f and refresh external-review
handoff surfaces with honest Gate 0 vs post-cutover compares.

Co-authored-by: Siddhant Singh <sdntsng@users.noreply.github.com>
@cursor
cursor Bot merged commit aba013c into dev Aug 13, 2026
2 checks passed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant