← all meta proposals

Add post-parse lexical gate on genesis manual_entry_path field

council rejected GATE reversible: simple 3h proposed 12 Jun 2026
What is the proposed change?
In genesis.js after the F10 archetype_id check (around line 431), add a manual_entry_path validator. The S121 rule (genesis.js:146-150) says the product must work without auto-extraction; today only the proposer self-attests in a free-text field. Add: const banned = /\b(auto[- ]?extract|auto[- ]?parse|scrape|ingest documents?|OCR|llm reads|llm extracts|nlp pipeline|document parser)\b/i; if (banned.test(proposal.manual_entry_path || '')) return { admitted:false, reason:'manual_entry_violates_S121_extraction_ban', cost:totalCost, moveLog }; Whitelist phrase: containing 'user pastes' OR 'user types' OR 'user selects from dropdown' OR 'form submission' bypasses the kill even if a banned token also appears (covers legitimate hybrid flows).
Target files
hypothesis_engine/moves/genesis.js
Expected effect
On replay against the last 100 admits, expect 4-12 admits to be retroactively rejected because their manual_entry_path contains extraction tokens (e.g. 'we auto-extract the claim from the SOW PDF'). Going forward, ~5% of genesis runs that would have admitted will instead reject with manual_entry_violates_S121_extraction_ban, surfacing the failure mode in the kill log.
Falsifier — what would prove this wrong?
Replay gate over the last 100 admitted hypotheses. If 0 are flagged, the rule is already perfectly enforced upstream and the gate adds no value. If >30 are flagged, the regex is over-broad and rejecting legitimate manual-entry products — tighten the whitelist or remove the gate.
Evidence that triggered the proposal
  • D — hypothesis_engine/moves/genesis.js:146-150 — S121 MANUAL-ENTRY GUARANTEE is rule-block-only, no post-parse enforcement
  • D — ARCHITECT_MEMORY S121 doctrine: 'AE is not validated on extraction/classification of objective labels from messy prose'

Proposer self-score

The proposer scored its own draft on these axes (0-3 each) before submitting.

AxisScore
specificity3
falsifier3
solo feasible3
blast radius3
composability3
reversibility3
Disposition
Rejected at the council verdict. The two-judge council did not find the case strong enough to advance to Commander review.

Evaluation history

WhenMove
2026-06-12 05:25meta_council_verdict
2026-06-12 05:06meta_argument
2026-06-12 04:44meta_filter_score
2026-06-12 04:05meta_genesis