Add reversibility-cost validator tool for proposed_diff

filter rejected TOOL reversible: simple 4h proposed 6 Jun 2026

What is the proposed change?

New tool reversibility_check.js: parses proposed_diff for keywords {schema migration, drop table, ALTER TABLE, rename column, breaking API, delete file, rm -rf, rewrite} and counts distinct target_files. Returns SIMPLE if 0 keywords and target_files<=3; MEDIUM if 1 keyword or 4-6 files; COMPLEX otherwise. filter_score.js calls it and overrides any self-declared reversibility that conflicts with the validator's assessment (logs override). Composite_v2 gets -2 if final reversibility is COMPLEX.

Target files

meta_engine/tools/reversibility_check.js meta_engine/moves/filter_score.js

Expected effect

Over 7 cycles, ~10-20% of proposals get reversibility downgraded by the validator (LLM self-reports SIMPLE more often than warranted). Graduated COMPLEX-reversibility proposal share drops from current visible rate to <=5%.

Falsifier — what would prove this wrong?

If the validator's reversibility assessment agrees with the LLM self-report in >95% of cases over 7 cycles, the tool adds no signal and is redundant — revert.

Evidence that triggered the proposal

D — brain/DESIGN_PRINCIPLES.md (reversibility-first)
D — brain/S158_FILTER_SCORE_AXES_v2.md (blast_radius axis)

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	3
reversibility	3

Disposition

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

Evaluation history

When	Move
2026-06-12 04:21	meta_filter_score
2026-06-06 04:03	meta_genesis