📊 Aggregating all the Agent Rubrics for review (Project Mirror v2) - #93
Draft
sugaroverflow wants to merge 8 commits into
Draft
📊 Aggregating all the Agent Rubrics for review (Project Mirror v2)#93sugaroverflow wants to merge 8 commits into
sugaroverflow wants to merge 8 commits into
Conversation
Introduce aggregate-mirror-v2-criteria.ts to pull Part A/B/C (or single constitution.md) from each project-mirror-v2/* ref and write committee-aggregation artifacts: prose review doc, long-format sections CSV, and per-fellow index CSV. npm run aggregate:mirror-v2-constitutions regenerates outputs after git fetch. Made-with: Cursor
sugaroverflow
requested review from
FrencHorses,
Gamithra,
alecsandrac,
asilsidahmed-Newspeak,
frederickobrien and
jcoombes
as code owners
March 29, 2026 14:30
Made-with: Cursor
sugaroverflow
changed the title
sugaroverflow
marked this pull request as draft
March 29, 2026 14:32
sugaroverflow
removed request for
FrencHorses,
Gamithra,
alecsandrac,
asilsidahmed-Newspeak,
frederickobrien and
jcoombes
March 29, 2026 14:32
… for split mirrors - Add Data tables section: fellow summary + part_a/part_b/part_c/full_constitution tables - Character counts align with sections CSV; Jump links to full text anchors - Index CSV: full_constitution_chars only for single_file layout (not duplicate file on split branches) Made-with: Cursor
Parse criteria from criteria.md, ## Part A in constitution.md, or markdown criterion tables; support ### Cn., ### 1., plain C1: lines, and | C1 | table rows. Add all-mirror-criteria-table.csv and Catalogue section in markdown (master + per-fellow tables). Document in README and PR body. Made-with: Cursor
… fellow) Drop mega markdown, char-count tables, SVG, and inlined full constitution text. Default output is ~300 lines: ## Name + | # | Criterion | Weight | per person. CSVs unchanged for full text and analysis. Made-with: Cursor
sugaroverflow
changed the title
- Part B: support Modifier N:, M1., M1 —, numbered ###, markdown tables, plain M1: lines, plain Direction/Magnitude lines, and title-based inference - Part C: require delimiter matches; H2 sections with **Rule:**/**Trigger:**; fix bold-colon detection for Trigger lines - Comparison stats: count conditional before boost/reduce - Refresh all-mirror-rubrics.md and pr-93-body.md Made-with: Cursor
Frederick O'Brien → Beacon, Martina Orlea → Signal, Connor Dunlop → Safeguard, Emily Mayhew → Harbour. Updates all-mirror-constitutions-index.csv, all-mirror-rubrics.md, all-mirror-criteria-table.csv, all-mirror-constitutions-sections.csv, and pr-93-body.md. PR links (#72, #80, #81) preserved as history.
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Project Mirror v2 — aggregate rubrics for review
Generated: 2026-03-29
Readable per-fellow + comparison tables live in all-mirror-rubrics.md (same content as below, without PR length limits).
This description is auto-generated — run
npm run aggregate:mirror-v2-constitutionsand paste orgh pr edit --body-file iterations/project-mirror-v2/committee-aggregation/pr-93-body.md.Pipeline
flowchart TB subgraph sources["18 fellow branches (project-mirror-v2/*)"] split["Most: criteria + modifiers + procedural"] single["Davit + Francesca: constitution.md only"] end run["npm run aggregate:mirror-v2-constitutions"] subgraph out["committee-aggregation/"] md["all-mirror-rubrics.md"] prb["pr-93-body.md — copy into GitHub PR"] crit["all-mirror-criteria-table.csv"] sec["all-mirror-constitutions-sections.csv"] end split --> run single --> run run --> md run --> prb run --> crit run --> secCross-fellow comparison
Overlap scores use Jaccard similarity on word tokens (length ≥ 4) from Part A criterion titles + Part B modifier titles only — a rough signal, not semantic equivalence.
Rubric shape
constitution.mdconstitution.mdModifier directions (Part B)
Theme signals in titles (Part A + B)
How many fellows have each theme substring anywhere in criterion or modifier titles (case-insensitive).
Most similar rubric wording (Jaccard on title tokens)
Least similar pairs (still some token overlap)
Full table — Part A (all fellows)
Master table — Part B value modifiers (all fellows)
Full table — Part C procedural rules (all fellows)
dossier_completeness< 0.25 with no usable description, AND (b) the project's websit…popularity_riskfield is set to HIGH for any project that: (a) has dossier_c…After fellow PRs change
Re-run the npm script and commit updated
all-mirror-rubrics.md,pr-93-body.md, and CSVs.