Improve homepage onboarding and document answerability results

content/docs/changelog.mdx

@@ -14,6 +14,18 @@ breaking changes require a major bump.
 
 ## Unreleased
 
+### Memory evidence and hardening
+
+- Opt-in scoped fact evidence, partial answers and persistent conflict warnings;
+  a rebuildable schema-15 projection binds facts to sources and claim revisions.
+- Durable corrections, bounded delivery, conservative lifecycle decisions and
+  atomic replay/sync. Remote context honors the server's disabled reranker.
+- The [paired answerability study](/docs/memory-benchmark/answerability) retained
+  24/27 hits and reduced unwanted injections 15/18 → 3/18 on its new synthetic
+  fixture. Exact metadata matched 36/45 in both repeats; p95 was 376.6 → 386.6 ms.
+  Compound retrieval and unsupported-subject fallback remain incomplete, and
+  the guard remains disabled by default.
+
 ### Security
 
 - Updated transitive `h2` from 0.4.14 to 0.4.16, fixing

content/docs/how-kimetsu-works/configuration.mdx

@@ -5,6 +5,15 @@ description: "Every knob in project.toml, the off-switches, and the environment
 
 Every knob in `project.toml`, the off-switches, and the environment overrides.
 
+## Unreleased fact guard
+
+On builds containing structured fact evidence, `broker.explicit_fact_guard`
+defaults to `false`. Enable with `kimetsu config set broker.explicit_fact_guard true`.
+It adds scoped supported/missing/conflicting evidence without extra model calls;
+injected context still consumes tokens. See [results, storage behavior and
+limitations](/docs/memory-benchmark/answerability). This setting is unreleased
+and is deliberately separate from the published configuration example below.
+
 Project config lives in `<project>/project.toml`:
 
 ```toml
@@ -39,7 +48,7 @@ budget_run_cap_tokens = 8000  # per-run ceiling on injected tokens
 compress_capsules = true      # strip tags, cap at 3 sentences; ranking unaffected
 session_dedupe = true         # skip capsules already injected this session
 warm_start = true             # SessionStart injects the digest + episodic resume
-answer_grade_min_score = 0.92 # top capsule >= this gets a "Verified answer" prefix
+answer_grade_min_score = 0.92 # relevance marker; not independent verification
 proactive_prefetch = false    # opt-in trajectory-based pre-fetch at PreToolUse
 
 [storage]

content/docs/memory-benchmark/answerability.mdx

@@ -0,0 +1,83 @@
+---
+title: "Structured answerability (unreleased)"
+description: "Paired evidence-delivery measurements, configuration and remaining gaps for the opt-in fact guard."
+---
+
+This is **unreleased, opt-in work**, not a new overall BrainBench score or a
+claim about generated-answer accuracy. It adds source-bound configuration facts
+and reports which requested attributes are supported, missing or conflicting.
+
+## Paired results — September 7, 2026
+
+The campaign contains 688 observations over 299 distinct scenario/query cases.
+Both sides enable the fact guard; the baseline is the preceding answerability
+build and the candidate adds structured evidence. No positive-hit losses or
+execution errors were observed.
+
+| Fixture | Positive hits, old → new | Unwanted injections, old → new | Subsequent p95, old → new |
+|---|---|---|---|
+| Development: 210 queries | 169/197 → 169/197 | 7/13 → 7/13 | 1,111.7 → 1,094.0 ms |
+| Prior answerability: 44 queries | 24/24 → 24/24 | 0/20 → 0/20 | 409.1 → 380.7 ms |
+| New structured facts: 45 queries, two repeats | 24/27 → 24/27 | 15/18 → 3/18 | 376.6 → 386.6 ms |
+
+On the new fixture, **80% fewer unwanted injections** means 15 failures became
+3, or 83.3% → 16.7% of the 18 negative cases. Exact status, values, missing
+attributes and conflicts matched **36/45 cases in each repeat**: 33/42 direct
+fact questions and all three broad controls. Rankings and metadata meaning
+were stable across repeats.
+
+The new-fixture p95 increased about **2.7%**, and mean MCP result size increased
+from **613.5 to 651.2 bytes** (6.2%). Peak MCP working set remained about
+653.5 MiB. These short runs do not establish zero overhead or a statistically
+significant speed change.
+
+All runs used cached BGE-small embeddings, default thread settings and a 6,000
+output budget. Development used TinyBERT with cutoff 0.30; the other fixtures
+used the optional quantized mMARCO reranker with cutoff 0.55. No model default
+was promoted. The byte budget includes serialization and is not billed tokens.
+
+## What changes for an agent
+
+With `Orchid staging gateway port is 7319.` recorded, a recognized port-and-timeout
+question can return port evidence and identify timeout as missing. Production
+evidence cannot fill a staging request. Conflicting eligible values stay marked
+as conflicting even when a capsule cap or budget removes one source. Exact
+unit comparison treats `30 seconds` and `30000 ms` as equivalent.
+
+On a build containing this change, enable it with:
+
+```sh
+kimetsu config set broker.explicit_fact_guard true
+```
+
+The default is `false`; set it back to `false` to disable it. Published packages
+may not contain this unreleased feature. An older warm daemon must be restarted
+with the updated build to use its new behavior.
+
+Schema 15 maintains a rebuildable SQLite projection bound to claim revisions,
+source events, validity and exact evidence excerpts. Corrections and replay
+refresh the projection; the tagged agent record path is covered. Ordinary
+retrieval skips fact hydration when the guard is disabled, but writes still
+maintain the projection. No extra model calls are needed; evidence sent to an
+agent still consumes context tokens.
+
+## Remaining gaps and measurement limits
+
+Six compound-question cases per repeat omitted port evidence: three retrieved
+nothing and three retrieved retries alone. Hit@4 counts the latter as hits,
+which shows why retrieval hit rate is not complete answerability. Three cases
+using the subject word `Unknown` fell outside the conservative grammar and
+returned unrelated evidence through legacy fallback. These account for the
+remaining unwanted injections.
+
+The fixture is assistant-authored and repeats templates across three project
+names. It is not 45 independent task families, does not measure an LLM reader,
+and cannot certify corpus-wide conflict detection. No grammar, parameters or
+expectations changed after these results. Retrieval per requested attribute and
+safer unsupported-subject handling remain follow-up work.
+
+The tested implementation passed 1,470 workspace tests (six ignored), 132
+benchmark Rust tests, 18 Python tests and six release CLI/MCP probes.
+
+Source: [full report, frozen fixtures, raw observations and hashes](https://github.com/RodCor/kimetsu/blob/281c6df/docs/audits/2026-09-07-structured-facts.md).
+

content/docs/memory-benchmark/brainbench.mdx

@@ -24,6 +24,10 @@ Brain Quality Index of 80.0%**. By capability:
 
 ## v2.6.0 re-measurement (264 scenarios)
 
+For the newer, unreleased paired evidence-delivery work, see
+[Structured answerability](/docs/memory-benchmark/answerability). Its synthetic
+45-query result is separate from the historical overall indices on this page.
+
 Re-run on the semantic build, all 264 scenarios of the comprehensive set, none
 skipped:
 

content/docs/memory-benchmark/index.mdx

@@ -9,6 +9,10 @@ them rather than take our word for it.
 
 ## The headline numbers
 
+The [unreleased structured-answerability study](/docs/memory-benchmark/answerability)
+reports a separate paired experiment: unwanted injections 15/18 → 3/18, with
+24/27 positive hits retained. It does not replace the published scores below.
+
 | benchmark | result |
 |-----------|--------|
 | LoCoMo | **89.4%** (1,540 questions, LLM-judged accuracy) |

content/docs/memory-benchmark/meta.json

@@ -4,6 +4,7 @@
     "index",
     "retrieval-and-correctness",
     "brainbench",
+    "answerability",
     "longmemeval",
     "beam",
     "comparison"