Repository navigation
feat(backfill): opt-in consolidation guards: content scope, duplicate filter, sensitivity rejection (T-1794) - #305
Merged
Merged
Conversation
# Conflicts: # docs/generated/release-truth.json # docs/generated/release-truth.md
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
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
T-1794 (AO-7): measured on 2026-10-25 over 300 random extracted candidates, only 18 of the 157 "useful" ones were about MemoryMaster while every capture carried
project:memorymaster(scope is taken from the session cwd), 93 of 157 already had a similar claim in the database, and 3 carried sensitive material.Change, behind
MEMORYMASTER_BACKFILL_CONSOLIDATE_GUARDS=1(default off, behaviour unchanged when unset): before a candidate is admitted to consolidation, (a) its scope is derived from repo/alias signals in its content via the existing scope alias resolution, generic rules are classified and skipped, (b) normalized-text duplicates of existing authorized claims are skipped and counted (reusesnormalize_claim_text), (c) every candidate passes the ingest sensitivity scan and rejected text is never persisted. Counts land in the run summary.Evidence (lane-run outside the worker sandbox): tests/test_dreaming_backfill_consolidate_guards.py + _cli + _quality + tests/test_public_demo.py = 77 passed. Worker: ruff, release-truth --check, contract-witness passed. Not yet done: re-measurement of 100 candidates from guarded output (needs the flag enabled on a campaign run, gated by T-1729).
🤖 Generated with Claude Code