superlocalmemory 4.0.9 → 4.0.10

package/CHANGELOG.md

@@ -5,6 +5,81 @@ All notable changes to SuperLocalMemory will be documented in this file.
 The format is based on [Keep a Changelog](https://keepachangelog.com/en/1.0.0/),
 and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0.html).
 
+## [4.0.10] — Your memories, not the summarizer's
+
+### Fixed
+- **A background pass had been writing its own summaries into your memory, and
+  they answered your questions.** SuperLocalMemory groups related memories and
+  writes a short summary of each group, meant for the dashboard. Since June
+  those summaries were being stored as memories, where they competed with your
+  own notes — and won, because each one carried the combined subjects of
+  everything it summarised and so looked more connected than any single note you
+  had written. Asked what you were working on, a store could answer with the
+  summarizer apologising for having nothing to say. Summaries now live
+  separately and are shown on the dashboard only. They are never returned as
+  answers, by any route: recall, search, the recent list, pinned context, an
+  export, or the memory count.
+- **Summarising a group set aside the memories in it.** That made sense while
+  the summary stood in for them in your answers. It does not now, so it has
+  stopped — and upgrading brings back the memories it had already set aside.
+  Where a store had been running long enough for those memories to have faded on
+  their own, they are restored to whatever standing their own history warrants,
+  not promoted.
+- **Upgrading repairs an affected store by itself.** No commands, no files to
+  edit. A verified backup is taken first, the repair is safe to interrupt and
+  safe to repeat, and it reports in plain language what it changed. It also
+  keeps checking: if the problem ever reappears, the next start handles it.
+- **The setting that turns summarising off did nothing.** It has been
+  documented and ignored since it was added.
+- **A summary that is a refusal is no longer stored as one.** When a group has
+  nothing in common, a model does not fail — it answers in prose, and that prose
+  was indistinguishable from a summary to anything that only checked whether
+  text came back. Model output is now cleaned of chat framing and rejected if it
+  turns out to be about the request rather than about your memories. A summary
+  assembled from your own sentences is exempt, so a note that happens to be
+  phrased like a chatbot is not thrown away.
+- **A summary of a group could pull real memories into a later grouping pass
+  and set them aside with it.**
+- **Summarising an unchanged group repeated its work indefinitely**, including
+  calling a language model to regenerate text it already had. It now recognises
+  that nothing has changed and stops.
+- **An exported backup carried the summaries**, and importing it would have
+  turned them back into memories on a machine where nothing was wrong.
+- **A memory can now be dated to when it happened**, not just to when it was
+  written. Everything was previously stamped with its ingestion date, so a note
+  written today about something from March filed itself under today. A date that
+  cannot be understood is refused rather than quietly dropped.
+- **A dated memory is now findable by its date even when nothing in it was
+  recognised as a name.** Roughly a quarter of a typical store was reachable
+  only by being recent, rather than by being about the right time.
+- **Stored memories now record which conversation they came from.** Reading
+  memories worked this out; writing them did not, so almost nothing carried it —
+  which left the feature that spreads an answer across different conversations
+  with nothing to work with.
+
+### Added
+- **`slm doctor` and the dashboard now tell you whether your memories can
+  actually be found**, with the counts behind the answer: how many you have, how
+  many are searchable by meaning, and how many are being held back and why. This
+  was previously unknowable without writing your own queries against the
+  database — one machine ran for months with 44% of its memory unreachable while
+  every status line said it was healthy.
+- **A Knowledge Overview on the Summaries tab.** What your memory knows, how
+  many memories each summary covers, and the stretch of time it spans. Summaries
+  that came back empty are counted rather than shown, and near-identical ones
+  are collapsed, so the page reflects what is there instead of padding itself.
+
+### Changed
+- **Searching by meaning no longer spends its results on held-back summaries.**
+  On a store with a lot of them, a question close to their subject matter could
+  come back with nothing usable at all while appearing to have answered.
+- **The startup repair no longer reports itself finished while memories remain
+  unsearchable.** It says how many are left, why, and that it will retry — and
+  it does retry, instead of recording success and never looking again.
+- **Dashboard files now carry a version taken from their own contents**, so a
+  changed script or stylesheet is always served as a new address. The version
+  numbers were maintained by hand and had stopped matching the files they named.
+
 ## [4.0.9] — Findable the moment you write it
 
 ### Fixed

package/README.md

@@ -5,7 +5,7 @@
   </picture>
 </p>
 
-<h1 align="center">SuperLocalMemory V4.0.9</h1>
+<h1 align="center">SuperLocalMemory V4.0.10</h1>
 
 <h2 align="center">Rent the LLM. Own the memory.</h2>
 
@@ -27,12 +27,12 @@ guarantee here is stated as a falsifiable invariant, tested under an adversarial
 negative control, and shipped with the harness that regenerates the evidence:
 <code>python benchmark/run_all.py --trials 200 --output-dir results/</code>. What each experiment
 does <em>not</em> exercise is stated too.</p>
-<p align="center"><code>v4.0.9</code> — one control plane: <strong>SLM-Mesh</strong> peer coordination · multi-scope memory (personal / shared / global) · profiles · Cache · Compress · 7-layer retrieval · code graph · Entity Explorer · skill evolution · Modes A/B/C · GDPR retention &amp; audit chain · bounded loops — across CLI, MCP, dashboard, the <strong>Claude plugin</strong>, the <strong>Codex add-on</strong>, and documented IDE integrations.<br/>
+<p align="center"><code>v4.0.10</code> — one control plane: <strong>SLM-Mesh</strong> peer coordination · multi-scope memory (personal / shared / global) · profiles · Cache · Compress · 7-layer retrieval · code graph · Entity Explorer · skill evolution · Modes A/B/C · GDPR retention &amp; audit chain · bounded loops — across CLI, MCP, dashboard, the <strong>Claude plugin</strong>, the <strong>Codex add-on</strong>, and documented IDE integrations.<br/>
 Proxy: <code>slm wrap claude</code> &nbsp;·&nbsp; MCP: add <code>slm_compress</code> to your config &nbsp;·&nbsp; Skill: zero-config</p>
 <p align="center"><strong>Four public arXiv preprints</strong> · V4: <a href="https://arxiv.org/abs/2608.08253">arXiv:2608.08253</a> · companion archive: <a href="https://zenodo.org/records/21853302">Zenodo 21853302</a> (<a href="https://doi.org/10.5281/zenodo.21853302">DOI 10.5281/zenodo.21853302</a>) · prior preprints: <a href="https://arxiv.org/abs/2603.02240">2603.02240</a> · <a href="https://arxiv.org/abs/2603.14588">2603.14588</a> · <a href="https://arxiv.org/abs/2604.04514">2604.04514</a>.</p>
 
 <p align="center">
-  <a href="CHANGELOG.md"><img src="https://img.shields.io/badge/v4.0.9-Current_Release-2ea44f?style=for-the-badge&logo=checkmarx&logoColor=white" alt="v4.0.9 — Current Release"/></a>
+  <a href="CHANGELOG.md"><img src="https://img.shields.io/badge/v4.0.10-Current_Release-2ea44f?style=for-the-badge&logo=checkmarx&logoColor=white" alt="v4.0.10 — Current Release"/></a>
   <a href="https://arxiv.org/abs/2608.08253"><img src="https://img.shields.io/badge/arXiv-2608.08253-b31b1b?style=for-the-badge&logo=arxiv&logoColor=white" alt="SuperLocalMemory 4.0 paper on arXiv:2608.08253"/></a>
   <a href="https://zenodo.org/records/21853302"><img src="https://img.shields.io/badge/Zenodo-10.5281%2Fzenodo.21853302-1682D4?style=for-the-badge&logo=zenodo&logoColor=white" alt="V4 paper on Zenodo: 10.5281/zenodo.21853302"/></a>
   <a href="https://arxiv.org/abs/2603.14588"><img src="https://img.shields.io/badge/arXiv-2603.14588-b31b1b?style=for-the-badge&logo=arxiv&logoColor=white" alt="arXiv Paper"/></a>

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "superlocalmemory",
-  "version": "4.0.9",
+  "version": "4.0.10",
   "description": "Local-first agent memory with MCP and an agent-native CLI. Documented clients include Claude Code, Cursor, and Windsurf.",
   "keywords": [
     "ai-memory",

package/plugin/.claude-plugin/plugin.json

@@ -15,5 +15,5 @@
   "mcpServers": "./.mcp.json",
   "name": "superlocalmemory",
   "repository": "https://github.com/qualixar/superlocalmemory",
-  "version": "4.0.9"
+  "version": "4.0.10"
 }

package/plugin/CLAUDE.md

@@ -1,4 +1,4 @@
-<!-- BEGIN SuperLocalMemory v4.0.9 -->
+<!-- BEGIN SuperLocalMemory v4.0.10 -->
 
 ## SuperLocalMemory (SLM) — Agent Rules
 
@@ -39,6 +39,6 @@ slm-recall · slm-remember · slm-session · slm-status · slm-cache · slm-comp
 ### Subagents
 slm-memory-advisor (memory decisions, session hygiene, scope/profile guidance) · slm-optimize-advisor (context compression + KV cache) · slm-governance-advisor (scope/roles/compliance/GDPR)
 
-<!-- END SuperLocalMemory v4.0.9 -->
+<!-- END SuperLocalMemory v4.0.10 -->
 
-SuperLocalMemory v4.0.9 · Qualixar · AGPL-3.0-or-later
+SuperLocalMemory v4.0.10 · Qualixar · AGPL-3.0-or-later

package/plugin/agents/slm-governance-advisor.md

@@ -77,4 +77,4 @@ slm-scope · slm-governance · slm-profile · slm-remember · slm-recall
 # What NOT to do
 Never session_init twice; never forget without dry-run preview; never store secrets; never bypass role checks; never claim an erasure succeeded without verifying via recall.
 
-SuperLocalMemory v4.0.9 · Qualixar · AGPL-3.0-or-later
+SuperLocalMemory v4.0.10 · Qualixar · AGPL-3.0-or-later

package/plugin/agents/slm-loop-runner.md

@@ -68,4 +68,4 @@ assessment. The gate is the authority.
 
 ---
 
-SuperLocalMemory v4.0.9 · Qualixar · AGPL-3.0-or-later
+SuperLocalMemory v4.0.10 · Qualixar · AGPL-3.0-or-later

package/plugin/agents/slm-memory-advisor.md

@@ -46,4 +46,4 @@ slm-recall · slm-remember · slm-session · slm-scope · slm-profile · slm-gov
 # What NOT to do
 Never session_init twice; never forget dry_run=False without reporting preview; never dump a whole file into remember; never invent a memory; never claim "saved" without success:true / clean CLI exit; never bypass scope or governance restrictions.
 
-SuperLocalMemory v4.0.9 · Qualixar · AGPL-3.0-or-later
+SuperLocalMemory v4.0.10 · Qualixar · AGPL-3.0-or-later

package/plugin/agents/slm-optimize-advisor.md

@@ -41,4 +41,4 @@ slm-compress · slm-cache · slm-status · slm-profile
 # What NOT to do
 Never compress code-for-edit/JSON-to-parse/<500 chars; never store secrets/ccr_ids; never let optimize failure block/alter the task; never claim a specific savings %; never carry ccr_ids across profile switches.
 
-SuperLocalMemory v4.0.9 · Qualixar · AGPL-3.0-or-later
+SuperLocalMemory v4.0.10 · Qualixar · AGPL-3.0-or-later

package/plugin/requirements.txt

@@ -1 +1 @@
-superlocalmemory==4.0.9
+superlocalmemory==4.0.10

package/plugin/skills/slm-cache/SKILL.md

@@ -145,4 +145,4 @@ These subcommands control daemon-level cache settings. They do not read or write
 
 ---
 
-SuperLocalMemory v4.0.9 · Qualixar · AGPL-3.0-or-later
+SuperLocalMemory v4.0.10 · Qualixar · AGPL-3.0-or-later

package/plugin/skills/slm-compress/SKILL.md

@@ -147,4 +147,4 @@ Content over 1 MB (1 000 000 bytes UTF-8) is processed but `reversible` is force
 
 ---
 
-SuperLocalMemory v4.0.9 · Qualixar · AGPL-3.0-or-later
+SuperLocalMemory v4.0.10 · Qualixar · AGPL-3.0-or-later

package/plugin/skills/slm-governance/SKILL.md

@@ -245,4 +245,4 @@ Before running any destructive operation (`forget`, `compact_memories`):
 
 ---
 
-*SuperLocalMemory v4.0.9 · Qualixar · AGPL-3.0-or-later*
+*SuperLocalMemory v4.0.10 · Qualixar · AGPL-3.0-or-later*

package/plugin/skills/slm-graph/SKILL.md

@@ -312,4 +312,4 @@ profile. See `slm-profile` for the full profile switching workflow.
 
 ---
 
-SuperLocalMemory v4.0.9 · Qualixar · AGPL-3.0-or-later
+SuperLocalMemory v4.0.10 · Qualixar · AGPL-3.0-or-later

package/plugin/skills/slm-loop/SKILL.md

@@ -96,4 +96,4 @@ paused, name the approval needed; when errored, quote the short detail.
 
 ---
 
-SuperLocalMemory v4.0.9 · Qualixar · AGPL-3.0-or-later
+SuperLocalMemory v4.0.10 · Qualixar · AGPL-3.0-or-later

package/plugin/skills/slm-mesh/SKILL.md

@@ -279,4 +279,4 @@ mesh availability.
 
 ---
 
-*SuperLocalMemory v4.0.9 · Qualixar · AGPL-3.0-or-later*
+*SuperLocalMemory v4.0.10 · Qualixar · AGPL-3.0-or-later*

package/plugin/skills/slm-profile/SKILL.md

@@ -146,4 +146,4 @@ Name them differently in your MCP config (e.g. `superlocalmemory-personal` and
 
 ---
 
-*SuperLocalMemory v4.0.9 · Qualixar · AGPL-3.0-or-later*
+*SuperLocalMemory v4.0.10 · Qualixar · AGPL-3.0-or-later*

package/plugin/skills/slm-recall/SKILL.md

@@ -236,4 +236,4 @@ before recalling, then switch back. See `slm-profile` for workspace switching.
 
 ---
 
-*SuperLocalMemory v4.0.9 · Qualixar · AGPL-3.0-or-later*
+*SuperLocalMemory v4.0.10 · Qualixar · AGPL-3.0-or-later*

package/plugin/skills/slm-remember/SKILL.md

@@ -238,4 +238,4 @@ different workspace, use `switch_profile` first. See `slm-profile`.
 
 ---
 
-*SuperLocalMemory v4.0.9 · Qualixar · AGPL-3.0-or-later*
+*SuperLocalMemory v4.0.10 · Qualixar · AGPL-3.0-or-later*

package/plugin/skills/slm-scope/SKILL.md

@@ -173,4 +173,4 @@ to review the impact. See `slm-remember` for the full deletion discipline.
 
 ---
 
-*SuperLocalMemory v4.0.9 · Qualixar · AGPL-3.0-or-later*
+*SuperLocalMemory v4.0.10 · Qualixar · AGPL-3.0-or-later*

package/plugin/skills/slm-session/SKILL.md

@@ -227,4 +227,4 @@ explicitly and call `recall` with `include_global`/`include_shared` after
 
 ---
 
-*SuperLocalMemory v4.0.9 · Qualixar · AGPL-3.0-or-later*
+*SuperLocalMemory v4.0.10 · Qualixar · AGPL-3.0-or-later*

package/plugin/skills/slm-status/SKILL.md

@@ -163,4 +163,4 @@ multi-profile setup. To switch the active profile, see `slm-profile`.
 
 ---
 
-SuperLocalMemory v4.0.9 · Qualixar · AGPL-3.0-or-later
+SuperLocalMemory v4.0.10 · Qualixar · AGPL-3.0-or-later

package/plugin-src/rules/AGENTS.md

@@ -129,4 +129,4 @@ When the SLM MCP server is unavailable, use these CLI equivalents:
 - **slm-optimize-advisor** — context compression and KV cache
 - **slm-governance-advisor** — scope/role compliance, retention policies, GDPR
 
-SuperLocalMemory v4.0.9 · Qualixar · AGPL-3.0-or-later
+SuperLocalMemory v4.0.10 · Qualixar · AGPL-3.0-or-later

package/pyproject.toml

@@ -1,6 +1,6 @@
 [project]
 name = "superlocalmemory"
-version = "4.0.9"
+version = "4.0.10"
 description = "Local-first agent memory with auditable hybrid retrieval"
 readme = "README.md"
 license = "AGPL-3.0-or-later"

package/src/superlocalmemory/__init__.py

@@ -32,7 +32,7 @@ if "OMP_NUM_THREADS" not in os.environ:
     os.environ["OMP_NUM_THREADS"] = "2"
 # ---------------------------------------------------------------------------
 
-__version__ = "4.0.9"
+__version__ = "4.0.10"
 
 _REQUIRED_VERSIONS = {
     "sentence_transformers": "5.3.0",

package/src/superlocalmemory/cli/commands.py

@@ -271,7 +271,7 @@ def _cmd_loop(args: Namespace) -> None:
 
 
 def _cmd_ops(args: Namespace) -> None:
-    """Wave-3: operational recovery & admin remediation commands."""
+    """Operational recovery & admin remediation commands."""
     from superlocalmemory.cli.ops_cmd import cmd_ops
     cmd_ops(args)
 
@@ -416,7 +416,7 @@ def dispatch(args: Namespace) -> None:
         "loop": _cmd_loop,
         # V3.8.2 super-help — grouped overview of every command + topics
         "help": cmd_help,
-        # Wave-3: operational recovery & admin remediation
+        # Operational recovery & admin remediation
         "ops": _cmd_ops,
         # V4.0.6: GDPR subject-rights CLI (Art.15/17/20)
         "gdpr": _cmd_gdpr_dispatch,
@@ -2949,6 +2949,48 @@ def cmd_doctor(args: Namespace) -> None:
                 f"enabled [{_surface_str}] {_stats}",
             )
 
+    # Memory answer-ability.
+    #
+    # Every other check here asks whether a component is installed. None of
+    # them asked the only question that matters to the owner: can my memories
+    # actually be found? One machine ran for months with 43.7% of its store
+    # unreachable while doctor reported everything green, because the reachable
+    # share was never counted.
+    memory_health_data: dict | None = None
+    try:
+        from superlocalmemory.core.config import SLMConfig as _HealthCfg
+        from superlocalmemory.core.memory_health import describe, measure
+        _h = measure(_HealthCfg.load().db_path)
+        memory_health_data = {
+            "live_facts": _h.live_facts,
+            "findable_by_meaning": _h.findable_by_meaning,
+            "missing_vector": _h.missing_vector,
+            "withheld_summaries": _h.withheld_summaries,
+            "display_summaries": _h.display_summaries,
+            "hidden_by_forgetting": _h.hidden_by_forgetting,
+            "inconsistently_hidden": _h.inconsistently_hidden,
+            "reachability": round(_h.reachability, 4),
+            "healthy": _h.healthy,
+            "unavailable": list(_h.unavailable),
+            "summary": describe(_h),
+        }
+        _detail = " ".join(describe(_h))
+        if _h.healthy:
+            _check("Memory answer-ability", "PASS", _detail)
+        elif _h.inconsistently_hidden or _h.reachability < 0.9:
+            _check(
+                "Memory answer-ability", "FAIL", _detail,
+                fix="slm restart",
+            )
+        else:
+            _check("Memory answer-ability", "WARN", _detail,
+                   fix="slm db reembed --missing-only")
+    except Exception as _mh_exc:  # noqa: BLE001 — a report must not break doctor
+        _check(
+            "Memory answer-ability", "WARN",
+            f"could not be measured: {_mh_exc}",
+        )
+
     # Summary
     if use_json:
         from superlocalmemory.cli.json_output import json_print
@@ -2959,6 +3001,7 @@ def cmd_doctor(args: Namespace) -> None:
         json_print("doctor", data={
             "checks": checks,
             "summary": {"passed": passed, "warned": warned, "failed": failed},
+            "memory_health": memory_health_data,
         }, next_actions=next_actions)
     else:
         print(f"\nSummary: {passed} passed, {warned} warnings, {failed} failed")

package/src/superlocalmemory/cli/main.py

@@ -977,7 +977,7 @@ def main() -> None:
 
     register_summary_parser(sub)
 
-    # Wave-3 / V4.0.6: GDPR subject-rights CLI (Art.15/17/20)
+    # V4.0.6: GDPR subject-rights CLI (Art.15/17/20)
     gdpr_p = sub.add_parser(
         "gdpr",
         help="GDPR subject rights: status, export (Art.15/20), erase (Art.17), verify",
@@ -1051,7 +1051,7 @@ def main() -> None:
         "--json", action="store_true", help="Output structured JSON",
     )
 
-    # Wave-3: operational recovery & admin remediation
+    # Operational recovery & admin remediation
     ops_p = sub.add_parser(
         "ops",
         help="Operational recovery: list failures, resolve stuck ops, check status",

package/src/superlocalmemory/code_graph/bridge/maintenance.py

@@ -103,6 +103,14 @@ def _fact_rows(
         # the pass linked happily in tests and would have linked NOTHING for any
         # actual user — 3,608 of 3,608 facts filtered out on this machine.
         sql += "AND COALESCE(archive_status, 'live') != 'archived' "
+    # Same reasoning for withheld rows: a model's non-answer has no code
+    # entities worth resolving, and a link from one would surface it in the
+    # bridge's enrichment text.
+    try:
+        if memory_db._has_quarantine_column():
+            sql += "AND COALESCE(quarantined, 0) = 0 "
+    except Exception:  # pragma: no cover - helper absent on an unusual manager
+        pass
 
     params: list[Any] = [profile_id]
     if since: