LORE 0.62.25 — Memory match errors

Memory match errors

• Fix memory replace|remove|move and filemap replace|remove reporting review_changed when --match hit no entry or several. They now return match_not_found or match_ambiguous and change nothing: the caller's match needs fixing, not a concurrent writer or stale review.
• Refuse a missing or empty --match, and a positional these commands do not take, as invalid_request. Previously memory replace "<old>" "<new>" joined both into the text and matched every entry; filemap remove <path> removed the only row of a one-row map.
• Report an unresolved or ambiguous project name (memory move --to, project move, backfill --project) the same way. review_changed remains for reviewed-snapshot, exact sync-replay and pending-listing changes.

LORE 0.62.24 — Diagnostic traces and guarded skill experiments

0.62.24 — 2026-10-19

Diagnostic traces

• Add trace capture|show for linked Claude/Codex actions and successful observations, with scrubbing, source hashes and explicit omission counts; automatic native review capture is opt-in.
• Keep traces local and separate from the reviewer digest. Provider success flags do not verify task outcomes.

Skill experiments

• Add skills evaluate for exact pending coding skills: freeze baseline, candidate, fixtures and verifier; require training improvement, validation non-regression and preservation of each replay success.
• Consume final tests once; retain immutable candidates, task traces and reports. Human approval remains required; the included deterministic fixture does not measure real model improvements.

Quality checks

• Check a 16 KiB skill cap, excessive growth and operator-supplied answer markers before execution; label changed proposals/baselines stale and damaged evidence unavailable in pending review.
• Scripts require human terminal authority and run without inherited credentials. They are trusted local code, with process deadlines and output limits; no sandbox or model adapter is supplied.

LORE 0.62.23 — background review history

0.62.23 — 2026-10-19

Background review history

• Fix worker::provider and run_claude saving background reviews as resumable user sessions; disable session persistence for Deriver, Dreamer and their shared review callers.
• Retain that setting during OAuth retries. Existing session files and Desktop imports require separate archival.

LORE 0.62.22 — Codex memory writes

0.62.22 — 2026-10-13

Codex memory writes

• Fix standalone::authority and detect_writer ignoring Codex session markers; interactive memory writes now apply instead of staging or receiving untrusted_write.
• Infer source_engine from Codex or Claude caller markers when LORE_ENGINE is unset; explicit overrides still win.
• Retain hook staging, detached-write gating, and MCP proposal review; empty session markers and engine labels grant no write authority.

LORE 0.62.21 — reviewed Python codegraph snapshots

0.62.21 — 2026-10-12

Reviewed Python codegraph snapshots

• Allow Python paths through codegraph_snapshot_read_v1 and codegraph_snapshot_store_v1, retaining exact-source rehash checks.
• Keep Python snapshot binding unknown: syntax candidates do not prove semantic imports, calls, or name resolution.

LORE 0.62.20 — reviewed codegraph snapshots

0.62.20 — 2026-10-09

Reviewed codegraph snapshots

• Add codegraph_snapshot_read_v1 for worktree-scoped Rust syntax snapshots. Reads recheck the requested source hash and return unknown binding; other-file candidates remain unverified.
• Add codegraph_snapshot_store_v1 and lore-rs codegraph store with terminal review and revision checks. Snapshots stay local; no TUI background write or sync occurs.

LORE 0.62.19 — Pending summary completeness

0.62.19 — 2026-10-09

Pending summary completeness

• Fix pending_for_sessions_v1 reporting complete: true and zero pending when pending/*.json has an invalid or non-UTF-8 filename stem. These entries now make every requested session incomplete; valid IDs remain visible.

LORE 0.62.18 — session-scoped pending summary

0.62.18 — 2026-10-09

Session-scoped pending review

• Add pending_for_sessions_v1 with source-project proposal IDs and explicit incomplete states for legacy, unreadable, overflow, or unresolved claims.
• Serialize pending reads with stage, resolve, sync, and project-move writes, so a complete result describes one namespace snapshot.

LORE 0.62.17 — flag internal sessions that start with an assistant row

0.62.17 — 2026-10-04

Internal sessions with an earlier assistant row

• Fix index::parse_record missing LORE's own provider sessions that start with an assistant row (the "Not logged in" first attempt): the prompt now flags the session when no earlier row is a user row and drops the rows already read.
• On a live store 161 such sessions were indexed as normal work after 0.62.14; lore index --force removes them.

LORE 0.62.16 — index refuses an over-limit transcript alone

0.62.16 — 2026-10-04

Index with an over-limit transcript

• Fix lore index aborting every run when one transcript exceeds a size limit (256 MB file, 8 MB line, 20000 messages): index::index now refuses that file alone and reports it as refused; other errors still abort.
• On a live store a 518 MB Codex rollout blocked the index after the 0.62.15 fix; the new test indexes a small session beside a transcript with a line past the limit.