Append full escalation transcript to non-convergent telemetry sink

filter rejected TOOL reversible: simple 3h proposed 20 May 2026

What is the proposed change?

In council_verdict.js, after the escalation block at lines 387-397 (where escalated: true is set), add a JSONL append call that writes the full round transcript (all 3 rounds of judge responses, the hypothesis text, and escalation metadata including hypothesis_id, timestamp, escalationReason) to meta_engine/data/non_convergent/YYYY-MM-DD.jsonl. Create the directory if absent. The write must be fire-and-forget (non-blocking, silent failure on write error) so it cannot break the main verdict path. No new DB columns, no schema change.

Target files

hypothesis_engine/moves/council_verdict.js meta_engine/data/non_convergent/

Expected effect

Every future council escalation produces a persisted JSONL record. After 30 days, the non_convergent/ directory contains records for ≥95% of escalation events as verified by comparing JSONL entry count to council_verdict escalated=true rows in engine.db.

Falsifier — what would prove this wrong?

If after 30 days the JSONL capture rate is below 85% of escalation events in engine.db (entry count mismatch), the append logic has a silent failure mode and must be debugged.

Evidence that triggered the proposal

Corpus E: council_verdict kill distribution — 2/9 recent verdicts required human decision (escalated=true), no structured full transcript available for post-hoc analysis or meta_engine Corpus E feed
Corpus D: META_ENGINE_S158_RED_TEAM_SYNTHESIS Survivor D — 'non-convergence telemetry sink, 3h, SIMPLE, approved pending implementation'
Corpus D: ARCHITECT_MEMORY session S157 — non-convergent transcripts identified as missing input for meta_engine self-improvement loop; gap prevents retrospective analysis of what made these hypotheses ambiguous

Proposer self-score

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

Axis	Score
specificity	3
falsifier	3
solo feasible	3
blast radius	3
composability	2
reversibility	3

Disposition

Rejected by filter_score. The proposal did not meet the bar for specificity, falsifiability, or solo-feasibility.

Evaluation history

When	Move
2026-05-20 04:11	meta_filter_score
2026-05-20 04:10	meta_genesis