superlocalmemory 4.0.7 → 4.0.8

package/CHANGELOG.md

@@ -5,6 +5,79 @@ 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).
 
+## [4.0.8] — Where you were actually looking
+
+### Added
+- **Summaries reach agents and the dashboard.** 4.0.7 gave the summary layer a
+  command; an agent still had no way to ask for one, and neither did anyone
+  reading the dashboard. Adds the `get_memory_summary` tool and a **Summaries**
+  tab under Memories with Today, Yesterday, and a picker listing the projects
+  SuperLocalMemory has actually recorded activity in. Every result states how
+  many memories it was built from and how much it covered.
+- **Bounded Loops has its own page**, alongside the other integrations. It shows
+  whether Bounded Loops is installed, which agreement the two products
+  negotiated, and which finished runs have been imported — and states plainly
+  that imported evidence is read-only: SuperLocalMemory never learns from it, no
+  file paths leave your workspace, and the tamper-evidence it carries is not the
+  same thing as independent verification.
+- **Mesh shows its history**, not only who is connected this second. Sessions
+  leave the mesh when they end, so the page was empty between sessions even on a
+  machine with plenty of mesh activity.
+
+### Changed
+- **Recall Lab is the first thing you see under Memories**, with example
+  questions to start from. Summaries sits next to it. The three browsing views
+  follow.
+
+### Fixed
+- **Memory stopped learning from use.** Recall marks which memories it returned
+  so that later use can be credited back, but the mark never reached the
+  assistant, so nothing was ever credited. Every recall was scored as neither
+  good nor bad, which is why recall quality sat at its starting value and no
+  source ever showed a settled quality signal however much the memory was used.
+- **The behavioural half of consolidation had never run.** Nothing triggered it,
+  and when it did run, two of its four steps failed without reporting anything.
+  SuperLocalMemory now runs it on a schedule while idle — never while you are
+  storing or recalling — and at session end.
+- **"This project" summaries could not work.** The button sent no project, and
+  the answer was always an error. SuperLocalMemory installs once for your whole
+  machine, so a browser tab has no way to know which project you mean; it now
+  lists the ones it has seen.
+- **Locally written summaries began with an apology.** The local model was never
+  told what it was writing, so it replied as if in a chat. It now returns the
+  summary and nothing else.
+- **A summary built from nothing claimed to cover everything.** A project with
+  recorded activity but no saved facts is now reported as a partial view.
+- **Multi-Agent Memory showed access tokens instead of agent names**, and because
+  one agent uses many tokens over time, a single agent appeared as dozens.
+- **Optimize invented a chart.** The cache hit-rate graph drew a rising trend
+  generated from the current value rather than measured; no history is recorded,
+  so it is gone until there is something real to plot. "Tokens saved" was
+  labelled as a monthly figure with an upward arrow that nothing ever set — it is
+  a total since install, and now says so. The page also explains why its numbers
+  are zero while its switches are on.
+- **One kind of memory pattern was never produced.** The query that finds it
+  asked for columns that do not exist, so it failed on every run and was skipped
+  silently.
+- **Code links never appeared for anyone.** Two separate reasons. The pass that
+  creates them matched no facts at all on a real store — it looked for an empty
+  archive marker, while live facts carry `live`, so every fact was excluded. And
+  the panel that displayed them was the search-results view, not the memory
+  drawer people actually open. Both fixed; on a real store this produces 7,470
+  links across 3,603 memories.
+- **"Atomic facts" always read "No atomic facts recorded for this memory."** The
+  drawer looked up child facts using the row's own id, but a row in that pane is
+  itself an atomic fact and its parent is a separate field. Records stored
+  directly, which have no parent, now say so rather than reporting missing data.
+- **Summaries ignored your mode.** The command never passed the configuration, so
+  every summary took the no-model path — a Mode B user got Mode A output. Ollama
+  now writes them when configured, falling back if it is unavailable.
+- **Summary highlights were unreadable.** Truncation counted characters but
+  ignored newlines, so one fact containing headings and blank lines rendered as
+  six lines and the list stopped being a list.
+- **"Active entities" listed internal identifiers** instead of names. It now
+  reads `Fixed (127), Gateway (124)`.
+
 ## [4.0.7] — Reachable
 
 Everything here was already built. None of it could be used.

package/README.md

@@ -5,7 +5,7 @@
   </picture>
 </p>
 
-<h1 align="center">SuperLocalMemory V4.0.7</h1>
+<h1 align="center">SuperLocalMemory V4.0.8</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>v4.0.7</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>v4.0.8</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/v4.0.7-Current_Release-2ea44f?style=for-the-badge&logo=checkmarx&logoColor=white" alt="v4.0.7 — Current Release"/></a>
+  <a href="CHANGELOG.md"><img src="https://img.shields.io/badge/v4.0.8-Current_Release-2ea44f?style=for-the-badge&logo=checkmarx&logoColor=white" alt="v4.0.8 — 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": "4.0.7",
+  "version": "4.0.8",
   "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": "4.0.7"
+  "version": "4.0.8"
 }

package/plugin/CLAUDE.md

@@ -1,4 +1,4 @@
-<!-- BEGIN SuperLocalMemory v4.0.7 -->
+<!-- BEGIN SuperLocalMemory v4.0.8 -->
 
 ## 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 v4.0.7 -->
+<!-- END SuperLocalMemory v4.0.8 -->
 
-SuperLocalMemory v4.0.7 · Qualixar · AGPL-3.0-or-later
+SuperLocalMemory v4.0.8 · 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 v4.0.7 · Qualixar · AGPL-3.0-or-later
+SuperLocalMemory v4.0.8 · Qualixar · AGPL-3.0-or-later

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

@@ -68,4 +68,4 @@ assessment. The gate is the authority.
 
 ---
 
-SuperLocalMemory v4.0.7 · Qualixar · AGPL-3.0-or-later
+SuperLocalMemory v4.0.8 · 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 v4.0.7 · Qualixar · AGPL-3.0-or-later
+SuperLocalMemory v4.0.8 · 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 v4.0.7 · Qualixar · AGPL-3.0-or-later
+SuperLocalMemory v4.0.8 · Qualixar · AGPL-3.0-or-later

package/plugin/requirements.txt

@@ -1 +1 @@
-superlocalmemory==4.0.7
+superlocalmemory==4.0.8

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 v4.0.7 · Qualixar · AGPL-3.0-or-later
+SuperLocalMemory v4.0.8 · 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 v4.0.7 · Qualixar · AGPL-3.0-or-later
+SuperLocalMemory v4.0.8 · 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 v4.0.7 · Qualixar · AGPL-3.0-or-later*
+*SuperLocalMemory v4.0.8 · 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 v4.0.7 · Qualixar · AGPL-3.0-or-later
+SuperLocalMemory v4.0.8 · 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 v4.0.7 · Qualixar · AGPL-3.0-or-later
+SuperLocalMemory v4.0.8 · Qualixar · AGPL-3.0-or-later

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

@@ -279,4 +279,4 @@ mesh availability.
 
 ---
 
-*SuperLocalMemory v4.0.7 · Qualixar · AGPL-3.0-or-later*
+*SuperLocalMemory v4.0.8 · 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 v4.0.7 · Qualixar · AGPL-3.0-or-later*
+*SuperLocalMemory v4.0.8 · 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 v4.0.7 · Qualixar · AGPL-3.0-or-later*
+*SuperLocalMemory v4.0.8 · 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 v4.0.7 · Qualixar · AGPL-3.0-or-later*
+*SuperLocalMemory v4.0.8 · 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 v4.0.7 · Qualixar · AGPL-3.0-or-later*
+*SuperLocalMemory v4.0.8 · 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 v4.0.7 · Qualixar · AGPL-3.0-or-later*
+*SuperLocalMemory v4.0.8 · 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 v4.0.7 · Qualixar · AGPL-3.0-or-later
+SuperLocalMemory v4.0.8 · 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 v4.0.7 · Qualixar · AGPL-3.0-or-later
+SuperLocalMemory v4.0.8 · Qualixar · AGPL-3.0-or-later

package/pyproject.toml

@@ -1,6 +1,6 @@
 [project]
 name = "superlocalmemory"
-version = "4.0.7"
+version = "4.0.8"
 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__ = "4.0.7"
+__version__ = "4.0.8"
 
 _REQUIRED_VERSIONS = {
     "sentence_transformers": "5.3.0",

package/src/superlocalmemory/cli/summary_cmd.py

@@ -86,6 +86,25 @@ def _active_profile() -> str:
     return "default"
 
 
+def _load_config() -> Any:
+    """The active SLMConfig, so Mode B/C get LLM-written summaries.
+
+    Without this the generators receive ``config=None``, ``get_mode_str`` returns
+    ``"a"``, and the LLM branch is unreachable — every summary comes out
+    extractive no matter which mode the user runs. That is not a graceful
+    fallback, it is the enrichment path never being offered.
+
+    Returns None on any failure, which lands on the extractive path. That is the
+    right fallback: a deterministic summary beats an error.
+    """
+    try:
+        from superlocalmemory.core.config import SLMConfig
+
+        return SLMConfig.load()
+    except Exception:
+        return None
+
+
 def _emit(result: Any, as_json: bool) -> None:
     """Print a SummaryResult as JSON or as prose."""
     if as_json:
@@ -123,6 +142,7 @@ def cmd_summary(args: Namespace) -> None:
     as_json = bool(getattr(args, "json", False))
     profile = getattr(args, "profile", None) or _active_profile()
     db = _db_path()
+    cfg = _load_config()   # Mode B/C enrichment; None -> extractive
 
     if not db.exists():
         print(f"No memory database at {db}. Run `slm status` first.")
@@ -131,7 +151,7 @@ def cmd_summary(args: Namespace) -> None:
     if sub == "session":
         from superlocalmemory.summaries import generate_session_summary
 
-        _emit(generate_session_summary(db, args.session_id, profile), as_json)
+        _emit(generate_session_summary(db, args.session_id, profile, cfg), as_json)
         return
 
     if sub == "day":
@@ -142,14 +162,14 @@ def cmd_summary(args: Namespace) -> None:
             target = (date.today() - timedelta(days=1)).isoformat()
         elif target == "today":
             target = date.today().isoformat()
-        _emit(generate_daily_reflection(db, target, profile), as_json)
+        _emit(generate_daily_reflection(db, target, profile, cfg), as_json)
         return
 
     if sub == "project":
         from superlocalmemory.summaries import generate_project_work_log
 
         path = getattr(args, "path", None) or str(Path.cwd())
-        _emit(generate_project_work_log(db, path, profile), as_json)
+        _emit(generate_project_work_log(db, path, profile, cfg), as_json)
         return
 
     print("Usage: slm summary {session <id> | day [DATE] | project [PATH]}")

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

@@ -96,7 +96,13 @@ def _fact_rows(
     except Exception:  # pragma: no cover - helper absent on an unusual manager
         has_archive = False
     if has_archive:
-        sql += "AND (archive_status IS NULL OR archive_status = '') "
+        # COALESCE form, matching storage/database.py:759 and the learning
+        # modules. My first version tested `IS NULL OR = ''`, which excluded
+        # every fact on a real store: live facts are marked 'live', not blank.
+        # The synthetic fixtures I developed against left the column unset, so
+        # 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' "
 
     params: list[Any] = [profile_id]
     if since:

package/src/superlocalmemory/core/consolidation_engine.py

@@ -198,27 +198,26 @@ class ConsolidationEngine:
                     from superlocalmemory.parameterization.prompt_injector import PromptInjector
                     from superlocalmemory.parameterization.prompt_lifecycle import PromptLifecycleManager
                     from superlocalmemory.learning.behavioral import BehavioralPatternStore
-                    # cross_project and workflow_miner may not exist yet — stub them
-                    try:
-                        from superlocalmemory.parameterization.cross_project import CrossProjectAggregator
-                    except ImportError:
-                        class CrossProjectAggregator:
-                            def __init__(self, db): pass
-                            def get_preferences(self, *a, **kw): return {}
-                    try:
-                        from superlocalmemory.parameterization.workflow_miner import WorkflowMiner
-                    except ImportError:
-                        class WorkflowMiner:
-                            def __init__(self, db): pass
-                            def mine(self, *a, **kw): return []
+                    from superlocalmemory.parameterization.cross_project import CrossProjectAggregator
+                    from superlocalmemory.parameterization.workflow_miner import WorkflowMiner
                     from superlocalmemory.hooks.auto_parameterize import AutoParameterizeHook
                     from superlocalmemory.core.config import ParameterizationConfig
                     p_config = getattr(self._config, "parameterization", ParameterizationConfig())
                     from superlocalmemory.infra.data_root import state_path
                     learning_db = str(state_path("learning.db"))
                     beh_store = BehavioralPatternStore(learning_db)
-                    cross_proj = CrossProjectAggregator(self._db)
-                    wf_miner = WorkflowMiner(self._db)
+                    # Both take a DB PATH, not a DatabaseManager. Passing
+                    # self._db here made Path(db_path) raise "argument should be
+                    # a str or an os.PathLike object ... not 'DatabaseManager'",
+                    # which step 9 caught at debug level — so soft prompts were
+                    # never generated and nothing said why.
+                    #
+                    # The ImportError stubs these replaced were also dead: both
+                    # modules exist and re-export the canonical classes, so the
+                    # stub branch could never run and only served to hide the
+                    # real signatures.
+                    cross_proj = CrossProjectAggregator(learning_db)
+                    wf_miner = WorkflowMiner(learning_db)
                     extractor = PatternExtractor(self._db, beh_store, cross_proj, wf_miner, p_config)
                     generator = SoftPromptGenerator(p_config)
                     injector = PromptInjector(self._db, generator, p_config)

package/src/superlocalmemory/core/recall_worker.py

@@ -106,6 +106,10 @@ def _handle_recall(
         memory_map=memory_map,
         per_fact_max=getattr(_rc, "recall_per_fact_max_chars", 2400),
         total_max=getattr(_rc, "recall_total_max_chars", 12000),
+        # Option B: markers only on session-bearing recalls. A marker can
+        # only buy a learning signal when a pending_outcomes row exists
+        # to settle, and those exist only when session_id is present.
+        include_marker=bool(session_id),
     )
     return {
         "ok": True,

package/src/superlocalmemory/evolution/skill_evolver.py

@@ -140,9 +140,24 @@ class SkillEvolver:
         # automatically. Tests inject their own.
         if budget is None:
             slm_home = canonical_data_root()
+            # ``learning_db`` must be learning.db: migration M010 creates
+            # evolution_llm_cost_log there. Every production caller
+            # (consolidation_engine, routes/evolution, MCP, CLI) passes
+            # memory.db as db_path, so handing db_path straight to the budget
+            # pointed it at a database where that table does not exist. Every
+            # evolution cycle then died with "no such table:
+            # evolution_llm_cost_log" — swallowed at debug level in
+            # consolidation step 11, so skill evolution silently never ran.
+            #
+            # Redirect only for the exact production filename. Tests pass
+            # ":memory:" or their own tmp file and must keep the old behaviour,
+            # rather than having a learning.db invented beside them.
+            learning_db = Path(self._db_path)
+            if learning_db.name == "memory.db":
+                learning_db = learning_db.parent / "learning.db"
             budget = EvolutionBudget(
                 profile_id=profile_id,
-                learning_db=Path(self._db_path),
+                learning_db=learning_db,
                 lock_dir=slm_home,
             )
         self._budget = budget

package/src/superlocalmemory/hooks/hook_handlers.py

@@ -768,7 +768,29 @@ def _run_quiet(cmd: list[str], timeout: int = 5, postprocess=None) -> str:
 
 
 def _maybe_consolidate() -> None:
-    """Run cognitive consolidation if last run was >24h ago. Non-blocking."""
+    """Ask the daemon for a full consolidation if the last one was >24h ago.
+
+    Three things were wrong here before 4.0.8, and together they meant the
+    behavioural half of consolidation never ran on a real install:
+
+    1. **It ran the wrong pipeline.** ``slm consolidate --cognitive`` invokes
+       ``CognitiveConsolidator.run_pipeline()``. Behavioural assertion mining,
+       soft prompts, skill performance and skill evolution are steps 8-11 of
+       ``ConsolidationEngine.consolidate()`` — a different class entirely. On a
+       store with thousands of tool events, ``behavioral_assertions`` stayed at
+       zero rows while the miner, run once by hand, produced 9 immediately.
+    2. **It marked success before trying.** The 24h timestamp was written
+       *before* the subprocess launched, so a run that failed instantly still
+       bought a full day of silence. The marker is now written only after the
+       daemon confirms it scheduled the work.
+    3. **It could not fail visibly.** stdout and stderr went to DEVNULL and the
+       handler ended in a bare ``except: pass``, so nothing anywhere recorded a
+       failure. Errors now surface on stderr, where the hook runner logs them.
+
+    Non-blocking: the daemon schedules the pass and returns immediately, so
+    session end is never held up by a run that takes minutes. The daemon's own
+    periodic timer is the backstop for sessions where this hook never fires.
+    """
     try:
         last_ts = 0
         last_consolidation = _last_consolidation_path()
@@ -780,16 +802,21 @@ def _maybe_consolidate() -> None:
         if (now - last_ts) < 86400:  # 24 hours
             return
 
-        # Update timestamp FIRST to prevent concurrent runs
+        started = _daemon_post(
+            "/api/v3/consolidation/trigger",
+            {"lightweight": False, "background": True},
+        )
+        if not started:
+            # No marker written — the next session end retries. The daemon's
+            # consolidation lock makes a duplicate request a cheap no-op.
+            print(
+                "slm: consolidation could not be scheduled (daemon unreachable)",
+                file=sys.stderr,
+            )
+            return
+
         os.makedirs(os.path.dirname(last_consolidation), exist_ok=True)
         with open(last_consolidation, "w") as f:
             f.write(str(now))
-
-        # Run consolidation in background (don't block session end)
-        subprocess.Popen(
-            ["slm", "consolidate", "--cognitive"],
-            stdout=subprocess.DEVNULL,
-            stderr=subprocess.DEVNULL,
-        )
-    except Exception:
-        pass
+    except Exception as exc:
+        print(f"slm: consolidation trigger failed: {exc}", file=sys.stderr)

package/src/superlocalmemory/learning/pattern_miner.py

@@ -367,18 +367,23 @@ def _mine_channel_and_coretrieval(
                 gen += 1
 
         try:
+            # Column names must match learning.db: the table is
+            # (fact_id_a, fact_id_b, co_count). This query asked for
+            # (fact_a, fact_b, co_access_count) and therefore raised
+            # sqlite3.OperationalError on every run since it was written —
+            # caught below, logged at WARNING, and otherwise invisible. Net
+            # effect: 912 co-retrieval edges on a real store never produced a
+            # single pattern, and the Brain pane had one fewer signal with no
+            # indication anything was missing.
             coret_rows = learn_conn.execute(
-                "SELECT fact_a, fact_b, co_access_count "
+                "SELECT fact_id_a, fact_id_b, co_count "
                 "FROM co_retrieval_edges "
-                "WHERE profile_id = ? AND co_access_count >= 3 "
-                "ORDER BY co_access_count DESC LIMIT 20",
+                "WHERE profile_id = ? AND co_count >= 3 "
+                "ORDER BY co_count DESC LIMIT 20",
                 (profile_id,),
             ).fetchall()
             if coret_rows and not dry_run:
-                top_pair = (
-                    dict(coret_rows[0]).get("co_access_count", 0)
-                    if coret_rows else 0
-                )
+                top_pair = dict(coret_rows[0]).get("co_count", 0)
                 store.record_pattern(
                     profile_id=profile_id,
                     pattern_type="co_retrieval_clusters",

package/src/superlocalmemory/mcp/profiles.py

@@ -17,13 +17,17 @@ from __future__ import annotations
 # v3.6.14 WP-01: Named profile definitions
 # ---------------------------------------------------------------------------
 
-_PROFILE_CORE: frozenset[str] = frozenset({  # 16
+_PROFILE_CORE: frozenset[str] = frozenset({  # 17
     "remember", "recall", "search", "fetch", "list_recent", "update_memory", "forget",
     "session_init", "close_session",
     "slm_compress", "slm_retrieve", "slm_cache_set", "slm_cache_get", "slm_optimize_stats",
     # A client that can propose a correction must be able to inspect and
     # authenticate its review; otherwise the core lifecycle is incomplete.
     "review_correction", "list_corrections",
+    # v4.0.8: the readable summary layer (issue #113). In CORE because the
+    # natural caller is the agent holding the conversation — an assistant
+    # asked "what did I work on yesterday" should not need a power profile.
+    "get_memory_summary",
 })
 
 # Portable Brain evidence must reach the coding-host profile shipped by the
@@ -34,7 +38,7 @@ _PROFILE_BRAIN: frozenset[str] = frozenset({
     "observe_bounded_loop_evidence",
 })
 
-_PROFILE_CODE: frozenset[str] = _PROFILE_CORE | _PROFILE_BRAIN | frozenset({  # 31
+_PROFILE_CODE: frozenset[str] = _PROFILE_CORE | _PROFILE_BRAIN | frozenset({  # 32
     "build_code_graph", "get_blast_radius", "query_graph",
     "semantic_search_code", "get_review_context", "detect_changes",
     # switch_profile lets a plugin/IDE session change the active workspace over
@@ -64,8 +68,11 @@ _PROFILE_FULL: frozenset[str] = frozenset({
     "slm_compress", "slm_retrieve", "slm_cache_set", "slm_cache_get", "slm_optimize_stats",
     # v3.8.0: bounded-loop tools (CLI + /slm-loop command + MCP).
     "slm_loop_run", "slm_loop_history", "slm_loop_show",
+    # v4.0.8: readable summaries (#113). In core, so it must be in full too —
+    # full is asserted to be a superset of core.
+    "get_memory_summary",
     # prestage_context remains registered but deliberately raw-server-only.
-}) | _PROFILE_FULL_MESH  # 49
+}) | _PROFILE_FULL_MESH  # 50
 
 _PROFILE_POWER: frozenset[str] = _PROFILE_FULL | frozenset({  # 61
     "get_version", "get_mode", "health", "consistency_check", "recall_trace",

package/src/superlocalmemory/mcp/server.py

@@ -105,6 +105,11 @@ _ESSENTIAL_TOOLS: set[str] = {
     "observe_bounded_loop_evidence",
     # Update, review, and list form one core correction lifecycle.
     "review_correction", "list_corrections",
+    # v4.0.8 (#113): readable summaries. Present here as well as in the named
+    # profiles because this set is the FALLBACK surface — it must mirror
+    # ``full``, and a tool that ships in the smallest profile ("core") cannot be
+    # missing from the fallback without a client silently losing it.
+    "get_memory_summary",
     # Memory management (2)
     "forget", "run_maintenance",
     # NOTE: prestage_context IS registered (see register_prestage_tool below)
@@ -280,6 +285,8 @@ from superlocalmemory.mcp.tools_ops import register_ops_tools
 register_ops_tools(_target, get_engine)  # Wave-3: operational recovery & admin remediation
 from superlocalmemory.mcp.tools_brain import register_brain_tools
 register_brain_tools(_target, get_engine)  # v4.0.2 portable Brain receipts
+from superlocalmemory.mcp.tools_summaries import register_summary_tools
+register_summary_tools(_target, get_engine)  # v4.0.8 issue #113 summary reads
 from superlocalmemory.mcp.tools_context import register_prestage_tool