amicus 4.6.0 → 4.6.1

package/.claude-plugin/plugin.json

@@ -1,6 +1,6 @@
 {
   "name": "amicus",
-  "version": "4.6.0",
+  "version": "4.6.1",
   "description": "Multi-model LLM Council + parallel AI window for Claude Code. Run structured council reviews across Gemini, GPT, DeepSeek and more — or fork a conversation to any model and fold the results back.",
   "author": {
     "name": "Christian Wagner"

package/CHANGELOG.md

@@ -5,6 +5,48 @@ All notable changes to Amicus are documented here. Format follows
 
 ## [Unreleased]
 
+## [4.6.1] - 2026-08-03
+
+### Added
+
+- **The MCP channel finally hears about new versions** (spec 2026-08-03). The MCP server now
+  runs the update check at startup and appends one flavor-aware notice block to the first
+  successful tool result of each server process (once per session, latched); `amicus_guide`
+  carries an always-on update line, and one `[amicus] update available` line lands in the
+  client's MCP log on stderr. The instruction is chosen config-first (`npx -y amicus@latest`
+  registrations are told a restart suffices; cached/pinned npx copies get the re-point-or-
+  clear-cache hint in the unverified voice; global installs get `npm install -g amicus`, from
+  where #33's stale-version warning takes over). Words only — no auto-update over MCP, no
+  periodic re-check; `NO_UPDATE_NOTIFIER=1` still disables the check entirely.
+- **A lost Stage-1 seat gets one more chance (SL-2).** A council sub-wave that dies before
+  its legs exist, or a leg that ends with no usable output, is relaunched exactly once —
+  serially, after the surviving launches settle. Recovery announces in the one voice
+  (`Recovered: seat X reviewed on retry — …`, a `stage1-retry` heal on
+  `run.json`/`verdict.json` `degrades[]`) and the run stays exit 0; a seat still dead after
+  its retry degrades exactly as before, with both attempts named in the why. Unconditional;
+  gated on the run's `--max-cost` position (an over-budget run skips the retry and records
+  the loss byte-identically to v4.6.0). Retry legs and their spend-ledger rows carry
+  `retryOfWaveId`. A healed critic counts as seated in `verdict.seatLoss`.
+
+### Changed
+
+- **The shipped second-opinion MODEL-NOTES seed was corrected and enriched** (owner-ruled
+  fold-back, PR #93): the haiku "hard-404" warning re-caused to the `ANTHROPIC_BASE_URL` `/v1`
+  convention split, GLM's stale reliability caution withdrawn on the v4.4.1 fence-extractor
+  replay evidence, pre-degrade-era claims re-grounded in the current announce/retry contract,
+  and three model sections plus the "peer consensus ≠ evidence on published numbers" rule added
+  from the field.
+
+### Removed
+
+- **The unused `rebuildElectron` remediation hint** (owner ruling 2026-08-03, closing Plan 3's
+  queued hint-voice question). It had no live call site — `doctor --fix`, the in-place Electron
+  self-heal, is its stated convergence target — and its prose asserted unverified causes ("after
+  an ABI mismatch or partial unpack"). Absence-pinned in `tests/remediation-hints.test.js`; a
+  reintroduction must adopt the unverified-cause voice. The same ruling keeps
+  `sweepSessionIndexTmp`'s confident voice: its cause is definitional (an atomic-write tmp orphan
+  has no other producer), not a guess.
+
 ## [4.6.0] - 2026-08-02
 
 ### Added (v4.6 milestone — the degrade announcement invariant, plans 1-4)

package/README.md

@@ -441,7 +441,7 @@ $ amicus status demo123 --json
   "taskId": "demo123",
   "status": "complete",
   "elapsed": "5m 0s",
-  "version": "4.6.0",
+  "version": "4.6.1",
   "model": "google/gemini-2.5-flash",
   "phase": "terminal"
 }

package/docs/ROADMAP.md

@@ -7,12 +7,14 @@ major jump, gated on funding/cofounder. The observability arc is split so the **
 first (v4.3)** and the **Electron "Council Workspace" (v4.4)** rides on top of it. `--dry-run` cost
 preview dropped to the backlog.
 
-Amicus is at **v4.5.0** (tagged 2026-07-28). Each 4.x rev below leads with the benefit, not the
+Amicus is at **v4.6.0** (tagged 2026-08-02). Each 4.x rev below leads with the benefit, not the
 plumbing.
 
-**Status:** v4.0 through **v4.5.0** have **shipped** — everything down to the v4.6 heading is a
-record of what landed, not a plan. **v4.6 (composition + tagging + GUI ergonomics) is the next
-rev.** v5.0 remains forward-looking.
+**Status:** v4.0 through **v4.6.0** have **shipped** — everything down to the v4.7 heading is a
+record of what landed, not a plan. **v4.7 (composition + tagging + GUI ergonomics) is the next
+planned rev** — that scope carried the number v4.6 here until the degrade-announcement-invariant
+milestone took the v4.6.0 release (2026-08-02); renumbered, not descoped. v5.0 remains
+forward-looking.
 
 > 📁 **Reading this from an npm install?** Some references below point at working documents that
 > live in the git repository and are deliberately **not** in the published package — anything under
@@ -95,7 +97,7 @@ stops hiding. Design: `docs/superpowers/specs/2026-07-27-v4.5-save-and-share-des
   client best able to show it. *(S–M; the pieces exist — see the design notes below.)*
 - **Council policy packs + full run-profiles** (bench + lenses + options + briefing template, invoke by name) — B7/F5 *(M)*
 - **Briefing templates + library** (F9) *(S–M)* — the foundation packs reference; the `{{input}}`
-  chaining variable and the `critique`/`refine` built-ins arrive with v4.6
+  chaining variable and the `critique`/`refine` built-ins arrive with the composition rev (now v4.7)
 - **Ride-along fixes** — FR-1 (a failed council seat can render perpetually live), the FR-2 ruling,
   RN-1/RN-5/RN-11 Workspace renderer fixes, TST-3 real-CDP abort pass *(S each; dispositions for
   all 17 open items are tabled in the design doc's §8)*
@@ -103,7 +105,37 @@ stops hiding. Design: `docs/superpowers/specs/2026-07-27-v4.5-save-and-share-des
 > Why here: save/share velocity multipliers that only pay off once councils are a command (v4.0)
 > and observable (v4.3/v4.4); auto-open makes the v4.4 surface discoverable on its best client.
 
-## v4.6 — "Compose your councils" *(specced after v4.5 ships — anti-rot rule)*
+## v4.6 — "A loss announces itself" *(the degrade announcement invariant)* — ✅ SHIPPED v4.6.0, 2026-08-02
+**Benefit:** a council run can no longer degrade quietly — every loss states what was lost, why,
+and what it does to the run, in one voice, on every surface (stderr, `run.json`, `verdict.json`,
+the report, `doctor`). The north star made mechanical: a correct-but-silent degrade fails the bar
+as hard as a crash.
+- **The ten-channel degrade announcement contract** — every loss routes through one sink
+  (`src/council/run-degrade.js`, the only code allowed to flip `degraded.value`, enforced by a
+  source-scan invariant test) and lands with mandatory what/why/effect on every surface, including
+  the report's new **"What was lost"** section — #85 *(L)*
+- **`verdict.seatLoss` derived from the degrade records** (#84 — a dead critic *leg* finally flips
+  `criticSeated`; the v4.5.2 seatLoss suites passed byte-unedited) + **Stage-2 judge legs get
+  `runStats` cost rows** (#83 — per-leg attribution for ~38% of a run's spend that had none) *(M)*
+- **`doctor` speaks the vocabulary** — `doctor --json` gains additive `degrades[]`, `--fix` prints
+  `Recovered:` lines in the one voice, and the engine hints state causes as **unverified** instead
+  of asserting an antivirus guess *(M)*
+- **Workspace discoverability from the CLI** — `watch` usage names `--ui` (#80), a CLI council run
+  with Electron present prints how to open the live Workspace (#81), `watch --ui` against an
+  `--out-dir` run names its cause (#82), and the Stage-5 verdict rebuild preserves
+  `seatLoss`/`degrades[]` (#87) *(S each)*
+- **Deliberate behavior changes** — dropped preset members and shared-server acquisition failures
+  now exit degraded (2) on every transport; judge rows raise reported cost totals vs v4.5.x
+  (`runStats` consumers keying by model must exclude `role: 'judge'`).
+- **Docs** — the full record is `CHANGELOG.md` §4.6.0; spec
+  `docs/superpowers/specs/2026-08-01-degrade-announcement-invariant-design.md`; plans 1–4 under
+  `docs/superpowers/plans/` (`2026-08-0*-v4.6-degrade-invariant-plan-*.md`).
+> Why it jumped the queue (2026-08-01): the v4.5.x field reports showed the engine was not losing
+> legs (11 four-seat council runs on v4.5.4, 10 clean) — but when a seat *was* lost, nothing told
+> the user which one. That silent-degrade class was ruled a north-star violation and took the rev
+> number; the composition scope below moved to v4.7.
+
+## v4.7 — "Compose your councils" *(renumbered from v4.6, 2026-08-02 — spec + fresh plan at kickoff per the anti-rot rule)*
 **Benefit:** councils chain — generate → critique → refine with no manual copy-paste — and history
 becomes navigable.
 - **Composable/chained waves** (`--input-from <id>` / `--prompt-file -` pipe + per-source digests) —
@@ -115,8 +147,8 @@ becomes navigable.
   remainder of TST-7
 - **README + docs update** *(S)*
 > The 2026-07-19 combined spec (`2026-07-19-v4.5-policy-packs-composition-design.md`) holds the
-> approved chaining/tagging/F10 design detail and is the primary input to the v4.6 brainstorm; it
-> is NOT executed as-written — v4.6 gets its own spec + fresh plan once v4.5 ships.
+> approved chaining/tagging/F10 design detail and is the primary input to the v4.7 brainstorm; it
+> is NOT executed as-written — v4.7 gets its own spec + fresh plan at kickoff.
 
 ### Deferred out of v4.4.1 into v4.5 (2026-07-27)
 

package/docs/publishing.md

@@ -48,7 +48,7 @@ Provenance is implied under trusted publishing.
 
 Run top-to-bottom before `npm version`:
 
-1. **MODEL-NOTES fold-back:** diff the machine-local ledger (`~/.claude/skills/second-opinion/MODEL-NOTES.md`) against the shipped seed (`skills/second-opinion/MODEL-NOTES.md`); port durable, machine-independent lessons into the shipped file (merge/prune, keep it tight — no run-ledger numbers, those live in `amicus council stats`).
+1. **MODEL-NOTES fold-back:** diff the machine-local ledger (`~/.claude/skills/second-opinion/MODEL-NOTES.md`) against the shipped seed (`skills/second-opinion/MODEL-NOTES.md`); cherry-pick generalizable lessons from the local ledger into the shipped seed — **per-section judgment, both directions, never a bulk copy** (the divergence is two-directional: adopt seed-side corrections back into the local ledger too). Merge/prune, keep it tight — no run-ledger numbers, those live in `amicus council stats`. Standing practice per the owner's ruling, 2026-08-03: shipped = curated seed, local = lab notebook.
 2. `npm test` green; `npm run lint` clean.
 3. **Integration tier, both rails:**
    - `npm run test:integration` (keyless) — must be **0 failures**. This is the same thing the `integration` job in `ci.yml` runs on every push, so it should already be green.

package/docs/usage.md

@@ -443,7 +443,7 @@ $ amicus status demo123 --json
   "taskId": "demo123",
   "status": "complete",
   "elapsed": "5m 0s",
-  "version": "4.6.0",
+  "version": "4.6.1",
   "model": "google/gemini-2.5-flash",
   "phase": "terminal"
 }

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "amicus",
-  "version": "4.6.0",
+  "version": "4.6.1",
   "mcpName": "io.github.BourbonDog/amicus",
   "description": "Multi-model LLM Council + parallel AI window for Claude Code. Run structured council reviews across Gemini, GPT, DeepSeek and more — or fork a conversation to any model and fold the results back.",
   "keywords": [

package/skills/second-opinion/MODEL-NOTES.md

@@ -4,8 +4,10 @@ This file is the `second-opinion` skill's evolving memory of **how to actually d
 well**. Read it before Stage 0 (council selection and launch); update it, with the user's
 approval, at the end of each run (Stage 6). Keep it tight — merge and prune rather than append.
 
-_Last updated: 2026-07-26 (v4.4.0 fold-back: alias-resolution hygiene, `council run --run-id`,
-debate defense/re-vote waves now exercised, haiku and glm notes; see changelog)._
+_Last updated: 2026-08-03 (per-section fold-back, both directions — the haiku "hard-404"
+re-diagnosed as a `/v1`-less `ANTHROPIC_BASE_URL`, pre-degrade-era claims re-grounded in the
+announcement contract (`degrades[]`/`seatLoss`/exit 2/one Stage-1 retry), three model sections and
+the peer-consensus≠evidence rule upstreamed from the field ledger; see changelog)._
 
 ## Global operating rules (all models)
 - **Fast path:** `council run` applies `--agent Plan` / `--no-context` / `--summary-length
@@ -22,14 +24,33 @@ debate defense/re-vote waves now exercised, haiku and glm notes; see changelog).
   defined in SKILL.md Stage 0) — no size cap, no shell-quoting hazards. Never inline a briefing as
   a CLI argument.
 - **Run in the background (`run_in_background: true`); you're notified on completion. Don't poll.**
+- **A stale MCP server kills councils with a distinctive signature — zero output on every leg.**
+  If any amicus MCP tool result carries the "Amicus was upgraded on disk (running vX, on-disk vY).
+  Restart your MCP client" notice, do NOT launch councils through the MCP tools — relaunch via the
+  CLI (`amicus council run`, a fresh process) or restart the client first. A version-skewed server
+  has timed out an entire Stage-1 wave with literally zero output bytes on every leg; that
+  all-legs-zero-output shape is what distinguishes it from model/provider failures, which produce
+  partial output or errors.
 - **Read results from the JSON documents** (`--json`): a wave's `legs[].summary` / a run's
   `summary` is the model's output; `status`/`error`/`counts` are ground truth for failures. Never
   scrape stderr logs to judge success.
 - **Transient provider errors** (502s, connection drops): re-run the affected leg (solo
   `amicus start --json`, same briefing file) or the wave — see per-model notes for
   model-specific signals. Never present a half-finished run as an answer.
+- **Seat losses are announced, retried once, and exit-coded — read the surfaces, don't diff
+  finding counts.** Every lost seat/leg is announced in one voice on stderr and recorded in
+  `run.json.degrades[]`, `verdict.json.degrades[]`, the report's **"What was lost"** section, and
+  `verdict.seatLoss`, and the run exits degraded (2). A Stage-1 wave or leg that dies is relaunched exactly once (serially, after the
+  surviving launches settle; skipped when the run is already over `--max-cost`): a heal announces
+  as a `Recovered:` line and the run stays exit 0; a seat still dead after its retry is recorded
+  with both attempts named in the why. A dead route still doesn't stop a council — it shrinks it —
+  but the shrinkage is no longer silent: confirm the bench you paid for from
+  `seatLoss`/`degrades[]` after every run, because a shrunken bench weakens the tiers (fewer
+  corroborators per finding).
 - **Credentials:** keys live in `~/.config/amicus/.env`. The legacy `~/.config/sidecar/.env`
-  fallback was removed in v2.0.0 (see `docs/SHIMS.md`). Configure with `amicus setup`.
+  fallback was removed in v2.0.0 (see `docs/SHIMS.md`). Configure with `amicus setup`. If
+  `ANTHROPIC_BASE_URL` is set anywhere in the environment, it must carry its `/v1` suffix for
+  direct-Anthropic legs — see the haiku section for the full diagnosis.
 - **PowerShell `--models` quoting (Windows):** always quote comma-separated model lists —
   `--models "gemini,gpt,deepseek"`. Unquoted, PowerShell splits on commas and amicus receives one
   mangled alias → instant arg-parse failure. (Now baked into every SKILL.md example.)
@@ -40,18 +61,50 @@ debate defense/re-vote waves now exercised, haiku and glm notes; see changelog).
   gpt/deepseek/grok have handled 82k-word agentic reads; gemini(-flash) and kimi stalled (narrate-
   stall / 25-min timeout / poller "Incomplete"). Pre-select proven long-read models or inline the
   text for large-context models.
+- **Long agentic reads can "narrate-then-never-deliver" even on capable models that DO finish
+  reading — bake the anti-narration hardening into the FIRST attempt, not the retry.** Models have
+  done a genuine full chunked read (confirmed by realistic input-token counts) and then ended the
+  turn with only process narration ("Continuing…"), never producing the review + JSON: the read
+  completed, but the turn budget ran out before the switch from reading mode to writing mode.
+  Whenever the task requires many sequential file-read tool calls before writing, the briefing
+  needs an explicit preamble: *"this is headless with no follow-up turn; you must produce the
+  complete deliverable by the end of THIS response; if running low on room, stop reading early and
+  write from what you have rather than deliver nothing."*
+- **Multi-FILE agentic reads stub out like long single reads — even WITH the anti-narration
+  preamble baked in.** Weaker readers have returned narration stubs of roughly 60–100 chars on a
+  multi-file packet despite the standard preamble. Treat a many-file briefing as a long-read task:
+  expect stubs, retry solo once with the identical briefing, then substitute the model — or
+  pre-concatenate the packet into one file for weaker readers.
+- **`--agent Plan` can trigger a literal, hard "planning mode — cannot execute" refusal on some
+  models** (terse, near-zero-token responses), not just the intended read-only tool restriction —
+  and prompt wording alone has not fixed it; switching that model to `--agent Build` has, immediately.
+  On the manual path, Stages 2-3 never touch the source file (the briefing is self-contained plus a
+  no-tools preamble), so `--agent Build` is safe there; reserve `--agent Plan` for Stage-1 legs
+  that read the source, and fall back to Build per-model on a hard refusal.
 - **Debate mode rarely fires on correctness questions — design for it or expect a no-op.**
   `--debate` only engages findings the tally marked Contested or Disputed, and judges agree far
   more than expected: **2 disputes in 123 adjudications across four councils** (gemini/gpt/qwen).
   Three of those four runs produced zero debatable findings, so the round never ran. Bug hunts
   converge — a race condition either exists or it does not. If you want a rebuttal round, brief a
   genuine judgement call (architecture, tradeoffs, "is this over-engineered"), not a defect hunt.
+- **Peer consensus is not evidence on published, checkable numbers — verify against a primary
+  source before applying such a finding, especially from the Confirmed tier and findings that
+  survived debate.** In one paid council, 2 of 23 findings were factually wrong about published
+  specs and BOTH sat in the tiers that are supposed to signal reliability: one reached Confirmed
+  on two peer agreements, the other hardened after a debate round in which a correct dispute was
+  talked out of its objection. A rebuttal is a rhetorical performance — judges re-voting on a
+  persuasive defense have no more access to the spec sheet than they did the first time. Any
+  finding that asserts a published figure (spec, capacity, rating, dimension, date) gets a
+  verification pass before you act on it.
 - **Read spend from `run.json`, never from a provider credit balance.** `usage.cost.amount` is the
   run total; per-leg costs are in `runStats[].usage.cost`. Only OpenRouter-routed legs move the
   OpenRouter balance — gemini/gpt/anthropic bill directly against their own keys, so inferring cost
   from that balance under-reports it badly (observed: ~6x low). A 3-model bench + chair + debate is
   roughly **$0.60-0.80 per run**, not cents; budget `--max-cost` accordingly or the chair gets
-  skipped mid-run (exit 2, degraded) when the debate legs push the total past the ceiling.
+  skipped mid-run (exit 2, degraded) when the debate legs push the total past the ceiling. As of
+  v4.6, `runStats` also carries Stage-2 judge rows (judge-tagged), so totals read higher than
+  pre-4.6 runs for the same bench; anything keying `runStats` by model should exclude
+  `role: 'judge'`.
 - **Expect agreement inflation in Stage-2 adjudication.** The judge contract defines `agree` by
   worked example ("an 'I missed this — it's valid' counts as agree") but gives no example for
   `dispute` and no positive definition of `neutral`, while requiring a verdict on EVERY finding —
@@ -105,7 +158,9 @@ debate defense/re-vote waves now exercised, haiku and glm notes; see changelog).
   **rescope** mechanism rather than a defense — raisers overwhelmingly AMEND (downgrade an
   overstated severity, narrow a scope) rather than DEFEND, and amendments are re-confirmed on the
   re-vote. Expect high-amend / low-defend; the idea usually survives, the severity claim often does
-  not.
+  not. In ideation councils, **cross-lens convergence** (independent lenses proposing the same
+  idea) is the strongest priority signal — compute it yourself when clustering; it is
+  complementary to, not the same as, the tally's agree/dispute tiers.
 
 ## Per-model notes
 
@@ -118,61 +173,115 @@ debate defense/re-vote waves now exercised, haiku and glm notes; see changelog).
 - **Red-team:** takes an adversarial brief well — high variance by design; use when consensus risk is high.
 - **Blind self-votes are inconsistent** (self-#1 in some runs, self-last in others) → discount self-votes either way.
 - Alias has resolved to **flash** tiers: fast, shallowest coverage, yet a recurring sharp fact/consistency checker (it alone refuted a bench-wide date error). Cheap cross-check value.
+- **Asserts quantitative "corrections" (timings, capacities) with unearned confidence** — one such
+  "severely wrong" correction was disputed 3-0, including its own blind self-dispute. Cross-check
+  its numeric claims before weighting; its honest blind self-disputes are a useful calibration tell.
 - **Unreliable on long agentic reads** (see global rule) — inline the text or swap models for book-length material.
 
 ### DeepSeek  (`--model deepseek` → via OpenRouter)
 - **Strengths:** resilient; produces strong, well-structured, well-cited critical analysis. A good
   default reviewer and a proven chair.
 - **Quirk:** occasional transient 502 mid-run → re-run the leg.
-- Proven chair (5 clean chairings) — decisive, well-structured synthesis.
+- Proven chair (many clean chairings) — decisive, well-structured synthesis.
 - As a Stage-1 reviewer of human-facing documents it **over-escalates severity** (typos/tenure → "blocker"); discount its blocker labels against peers. Prune its self-retractions when tallying.
 - **"Agree-with-the-adversary" lean:** it has been the lone endorser of a red-team's harshest claims, turning them Contested — cross-check before treating its lone agreements as consensus.
+- **Stub-on-fanout / clean-on-solo-retry:** its Stage-1 fanout leg has returned a sub-100-char
+  narration stub despite the anti-narration preamble, then produced a full review on a plain solo
+  retry with the identical briefing — retry solo before substituting the model.
 
 ### GPT  (`--model gpt` → via OpenRouter)
 - **Strengths:** reachable via the OpenRouter key; resilient; very thorough structured critique
   (25 findings on a 1-page framework). Cleanly separates the review criteria.
 - **Quirks:** verbose — peers dinged it for volume-over-judgment (good coverage, lower
   discrimination); **self-ranked its own review #1** in cross-review → discount self-votes.
-- Ranked genuine #1 by all judges (incl. non-self) in one run — thoroughness is real, not a self-vote artifact. Handled an 82k-word agentic read cleanly.
+- Ranked genuine #1 by all judges (incl. non-self) in one run — thoroughness is real, not a self-vote artifact. Handled an 82k-word agentic read cleanly — but a later similar-size read delivered only narration until retried with the anti-narration preamble; apply the preamble to gpt by default, not just to weaker models.
 - **Asserts context-dependent facts (dates, "is this future?") without verifying** — and self-confirms them in adjudication. Cross-check any time-dependent claim it raises.
-- A good calibration anchor in cross-review: confirms observational findings, disputes interpretive overreach.
+- A good calibration anchor in cross-review: confirms observational findings, disputes interpretive overreach. A repeatable role: the seat most likely to catch **confidence inflation** in an otherwise-correct argument (theoretical risk asserted as demonstrated mechanism, proxies used as pseudo-diagnostics).
+- **When gpt disputes a specific numeric claim, weight that dispute heavily before a debate round erodes it** — it has been right and then talked out of the objection by a persuasive rebuttal (see the peer-consensus rule).
 
 ### Grok  (`--model grok` → via OpenRouter)
 - Very fast legs; credible judge and chair (rejected its own weak findings as chair; honest blind self-rank).
 - Strong red-team fit; handled an 82k-word agentic read. Weight its **observational** catches heavily and its **interpretive** verdicts cautiously (bench pattern: the former confirmed, the latter disputed).
 - Stage-1 non-red-team reviews skew to scope-inflated "missing content" majors.
+- Has **hard-refused under `--agent Plan`** ("Plan mode active… cannot execute"; near-zero tokens)
+  where `--agent Build` worked immediately — prefer Build for grok legs (see the global Plan-refusal rule).
+- **Fabrication risk on long-document tasks:** one fast, suspiciously low-fresh-token response
+  contained an invented "major"-severity finding plus unsupported flavor details, verified false
+  against the source. A later run under Build (genuine read, no fabrications on spot-check)
+  partially rebuilt confidence — usable for red-team slots under Build **with the spot-check
+  discipline**: verify its most specific/surprising claims against the source before letting them
+  into a council.
 
 ### Kimi  (`--model kimi` → via OpenRouter)
 - The bench's sharpest adjudicator (caught strawmen and misreads other judges waved through).
-- **Very slow legs (5-7 min)** — it gates wave wall-clock; budget timeouts around it.
-- Stalls on long agentic reads (poller "Incomplete" with only a preamble). Reserve for short-artifact adjudication.
+- Repeatedly the **top-ranked reviewer on consumer-practical, safety-heavy artifacts** (food
+  safety, mechanical advice) — top-ranked in two such runs (unanimous in one), with the most
+  granular findings on the bench and genuine structural catches.
+- Its specificity is also its risk: it reaches for **exhaustive quantitative claims ("every",
+  "all", "none") that are directionally right and literally wrong** — and it has successfully
+  DEFENDED one in a debate round against a correct dispute. Verify its universal quantifiers
+  against a primary source (see the peer-consensus rule).
+- **Very slow legs (4-8× its peers)** — it gates wave wall-clock; budget timeouts around it.
+- Stalls on long agentic reads (poller "Incomplete" with only a preamble). Reserve for short-artifact work.
 
 ### Mistral  (`--model mistral` → via OpenRouter)
 - Fast, broad coverage, catches real issues.
 - **Hallucination risk is real:** has invented non-existent product models/specs, disputed independently by two judges. Cross-check every specific model number or product claim it introduces.
 
 ### Claude  (in-council, when toggle on)
-- Consistently the most *calibrated* reviewer (no severity inflation; findings overwhelmingly Confirmed; bench-best street-cred in recent runs) but sometimes the least *original* — it can miss the boldest single catch. Treat as a reliability floor, not a discovery engine.
-
-### haiku  (`--model haiku`) — **verify before using; it has been hard-404ing**
-- The direct-Anthropic route (`anthropic/claude-haiku-4-5-20251001`) returned a hard `Not Found`
-  on **every** invocation of a recent paid corpus — 3 of 3 legs across two separate runs, both as
-  **chair** (twice, incl. the fallback retry) and as a **bench seat** — in ~2 s with zero tokens.
-- Both runs degraded *around* it silently rather than failing: one fell back to another chair, the
-  other collapsed its bench from 3 seats to 2 and exited 2. **A dead alias does not stop a council;
-  it shrinks it.** In the bench-seat case every finding came out `confidence: "thin"` with a single
-  peer corroborator, purely because the bench had halved — and nothing in `verdict.json` said so.
-- Resolve the alias against the catalog (`amicus models --check`) before putting it on a paid
-  bench, and re-check the bench roster in `run.json` against what you asked for afterwards.
-
-### GLM  (`--model glm` → z-ai via OpenRouter)
-- Cheap and fast, and ranked best-by-peers on a clean debut — but a later run produced **35 KB of
-  prose with `conformance: unstructured` and 0 parsed findings**, twice. Its structured-output
-  reliability is **not** established; treat the debut as low-N.
-- Useful behavioural note: when its structured output failed it **refused to fabricate** on the
-  repair attempts. Honest — but an honest refusal still costs you the seat, so a bench that leans
-  on `glm` for quorum can silently adjudicate a seat short while still paying for its tokens.
-  Watch `conformance` per seat, not just the finding count.
+- Consistently the most *calibrated* reviewer (no severity inflation; findings overwhelmingly Confirmed; bench-best street-cred in recent runs) but sometimes the least *original* — it can miss the boldest single catch. Treat as a reliability floor, not a discovery engine — though on some benches its findings have anchored the entire Confirmed tier.
+- Its own checkable, arithmetic-style claims deserve the same **mechanical verification** as everyone else's: one calendar-consistency claim in a Claude first-pass review was a genuine reasoning error, caught only by computing the dates programmatically before submission (see General).
+
+### haiku  (`--model haiku`) — **the "hard-404" was the environment, not the model**
+- **The model was never the problem.** The direct-Anthropic route returned an instant `Not Found`
+  (~2 s, zero tokens) on every invocation of a paid corpus — 3 of 3 legs across two runs, as chair
+  and as bench seat — and the standing diagnosis was a rotten alias. It is not: the cause is an
+  **`ANTHROPIC_BASE_URL` set in host form, without its `/v1` suffix**. Anthropic SDKs (including
+  Claude Code itself) treat the var as a HOST and append `/v1` themselves; OpenCode's provider
+  layer treats it as the full prefix — so a value that is correct for the host app kills every
+  direct-Anthropic leg amicus launches. Proven by control pair: an identical `fanout --models
+  opus` call fails "Not Found" on the host form and completes with `/v1` appended.
+- **Symptom signature:** instant "Not Found", zero tokens, direct-Anthropic routes only —
+  OpenRouter-routed legs in the same run are unaffected. It hits every direct-Anthropic alias
+  equally (haiku, opus, sonnet, claude; `fable` is OpenRouter-only and unaffected), so a "dead"
+  cheap seat and a "dead" frontier chair with this signature share one cause.
+- **The check:** inspect `ANTHROPIC_BASE_URL` in the environment amicus actually runs in — the var
+  can live only in a parent process's env (e.g. the host app), absent from every shell profile and
+  settings file on disk. If it lacks `/v1`, that is the kill. A doctor check for this is filed on
+  the backlog; until it ships, check by hand before blaming a model or an alias.
+- Seat-loss mechanics for a dead route are the same as any other loss — announced in one voice,
+  exit 2 if still dead. A dead Stage-1 seat gets the one retry; a dead CHAIR walks the chair's own
+  chain instead — same-chair retry, then promotion of the best non-bench model from the ledger —
+  with the actual chair checkpointed into `run.json` (see the seat-loss bullet in Global
+  operating rules).
+
+### GLM  (`--model glm` → `glm-5.1` via OpenRouter; the recent observations below are from the explicit `openrouter/z-ai/glm-5.2` id)
+- The v4.4.0-era "structured-output reliability not established" warning is **withdrawn as
+  wrong-cause**: its `unstructured`-conformance results and repair refusals traced to two
+  since-fixed engine defects (the unanchored fence extractor that truncated any JSON quoting a
+  code fence, and repair prompts that omitted the artifact under repair — both fixed in v4.4.1),
+  not to the model. Conformance has been `clean` in subsequent paid runs.
+- Cheap and fast; **ranked best-by-peers on an ideation-bench debut** — a focused,
+  fewest-findings reviewer whose findings land. Promising budget-to-mid bench member.
+- **First recorded confident fact error:** it asserted an engine spec that the primary source (the
+  owner's manual) contradicts outright — and the finding reached Confirmed on two peer agreements,
+  denied only by a post-council verification pass. Treat glm's confident factual assertions as
+  unverified until checked (the peer-consensus rule exists because of findings like this one).
+- Honest under pressure both ways: it refused to fabricate on repair attempts, and it has
+  withdrawn a contested finding cleanly in debate. An honest refusal still costs the seat — watch
+  `conformance` per seat, not just the finding count.
+
+### Qwen  (`--model qwen` → qwen3.7-max via OpenRouter; distinct from `qwen-coder`)
+- Very large context (1M tokens per catalog). As a red-team substitute it has produced a thorough,
+  well-organized adversarial review with accurate, specific line citations and genuinely unique
+  catches that verified true against the source — weight its specific, cited claims heavily.
+- Same observational-vs-interpretive split as grok: its cited observational findings get
+  confirmed; its "blocker"-severity interpretive claims get disputed as genre-normative. Weight
+  its severity labels on broad interpretive claims cautiously.
+- **Do not assign qwen Stage-1 multi-file reads.** Its proven mode is single-file reads and
+  short-artifact work: a book-length single-file read succeeded on retry with the anti-narration
+  preamble, but multi-file packets have produced narration stubs twice in one run — with the
+  preamble present, under both Plan and Build. It narrates rather than refuses under Plan.
 
 ### minimax  (`--model minimax` → via OpenRouter)
 - Fast (~2 min review legs), cheap, `clean` findings-JSON conformance on debut.
@@ -189,6 +298,21 @@ debate defense/re-vote waves now exercised, haiku and glm notes; see changelog).
 - Ranked last on its debut bench — misses the offline, interruption-handling, and
   test-methodology gap classes stronger seats catch. Fine budget-bench filler; do not chair it.
 
+### GPT-5.6-Terra  (`openrouter/openai/gpt-5.6-terra`)
+- Debut (one paid ideation run): `clean` conformance, thorough — the most findings on its bench.
+  Handled an inlined ~13K-token digest cleanly under `--agent Build`.
+- The priciest leg on its bench (~5× the cheap seats) and it self-ranked #1 (discount self-votes,
+  as always) while peers ranked it lower — thoroughness did not convert to peer standing. Low-N;
+  re-confirm before leaning on it.
+
+### Gemini 3.1 Pro  (`openrouter/google/gemini-3.1-pro-preview` — the PRO tier; the bare `gemini` alias resolves to flash)
+- Debut (one paid ideation run): `clean` conformance, mid-bench peer standing.
+  No narration or stall — but the material was INLINED, not an agentic read (where gemini-flash
+  historically stalls); don't extend the result to agentic reads untested.
+- Use the shipped `gemini-pro` alias when you want a real Pro reviewer — it live-resolves to the
+  current Pro tier, falling back to `openrouter/google/gemini-3.1-pro-preview`; the bare `gemini`
+  alias gives you flash.
+
 ### (others — add as used)
 - Opus / o-series etc. are reachable via amicus **if their API keys are configured**. Add notes
   here the first time each is used.
@@ -209,7 +333,7 @@ This section keeps only per-model **qualitative quirks** and **structural-confor
 
 - **deepseek** — strong synthesis, resilient; occasional transient 502 → re-run the leg. Proven chair. Conforms cleanly.
 - **gpt** — thorough but verbose; peers have dinged it for volume-over-judgment. Self-ranked its own review #1 in the 2026-06-04 run → the peers-only street-cred rule (now enforced by `tally`) mitigates this. Conforms cleanly. Accessible via OpenRouter.
-- **gemini** — fast, very large context; tends toward absolute severity labels ("blocker" inflation vs peers). Conforms cleanly; watch for preamble narration — instruct it to emit the JSON block verbatim after the prose.
+- **gemini** — fast, very large context; the bare alias resolves to flash tiers (shallow coverage, cheap fact-checks); asserts numeric "corrections" with unearned confidence — cross-check them. Conforms cleanly; watch for preamble narration — instruct it to emit the JSON block verbatim after the prose.
 
 ## Free-tier models (OpenRouter `:free`)
 - Heavily rate-limited (shared daily pool); a 3-leg parallel wave + cross-review can 429 mid-run.
@@ -229,6 +353,12 @@ This section keeps only per-model **qualitative quirks** and **structural-confor
 ## General
 - Model citations are usually real but **verify any load-bearing reference before publishing**;
   watch for loosely-attached attributions (e.g., a real paper cited for the wrong claim).
+- **Checkable, arithmetic-style claims (calendar/date consistency, counts, named specific facts)
+  deserve a mechanical verification pass before publishing — including claims from Claude's own
+  review, not just other models'.** One calendar-consistency claim in a Claude first-pass review
+  was wrong; running the actual dates through a real calendar caught it before it reached the
+  council or the author. This class of claim is cheap to verify mechanically and expensive to get
+  wrong in a document going to a human — don't re-reason it a second time; compute it.
 - Prefer models from **different families** for genuinely independent opinions.
 
 ## Lessons changelog
@@ -237,7 +367,7 @@ This section keeps only per-model **qualitative quirks** and **structural-confor
   its occasional transient 502.
 - **2026-06-03** — v2 council upgrade: added cross-review (Stage-2 anonymized peer ranking +
   per-finding adjudication) and reviewer-reliability tracking.
-- **2026-06-04** — Trusst messaging-framework council (Gemini + GPT + DeepSeek + Claude-in-council;
+- **2026-06-04** — B2B messaging-framework council (Gemini + GPT + DeepSeek + Claude-in-council;
   DeepSeek chair). First GPT use → per-model note. First scored reviewer-reliability rows
   (deepseek 2.33/100%, gpt 2.67/92%, gemini 3.67/89%; 1–4 scale).
 - **2026-06-10** — v3 migration: transport moved to `fanout --json` + `--prompt-file` (F4); pruned
@@ -254,11 +384,28 @@ This section keeps only per-model **qualitative quirks** and **structural-confor
   a bench seat; `council run` prints nothing until terminal, so pin `--run-id` when backgrounding.
   Debate's defense/re-vote waves exercised for the first time — they work, and behave as a rescope
   (amend-heavy) rather than a defense. New per-model notes: **haiku** hard-404ed 3/3 legs across two
-  runs and both councils silently degraded around it; **glm** returned `unstructured` conformance
+  runs and both councils silently degraded around it *(both halves since superseded: the 404 was a
+  `/v1`-less `ANTHROPIC_BASE_URL`, not the alias — see the haiku section — and the silent-degrade
+  era ended with the v4.5.2→v4.6 announcement contract)*; **glm** returned `unstructured` conformance
   with 0 findings twice after a clean debut, and honestly refused to fabricate on repair — which
-  still costs the seat.
+  still costs the seat *(since re-diagnosed: both results were v4.4.0 engine defects, fixed in
+  v4.4.1 — see the GLM section)*.
 - **2026-07-14 (v2.2.0)** — Optional council elements shipped and verified on a planted-flaw
   ground-truth council (critic seat, debate mode nothing-to-debate path, chair verdict scale;
   expert lenses defined but not yet field-run). New lessons: claim-class dedup glosses
   rationale-level errors in Stage-2; minimax debut (strong critic seat, CJK-intrusion quirk);
   qwen-coder debut (fast budget filler). Debate mode's defense/re-vote waves still unexercised.
+- **2026-08-03 (fold-back, both directions — owner's ruling)** — Per-section reconciliation with
+  the machine-local field ledger after six release-cycle deferrals. Corrections: the haiku
+  "hard-404" re-diagnosed (a `/v1`-less `ANTHROPIC_BASE_URL` in host form — the model was never
+  the problem; control pair: identical `fanout --models opus` fails host-form, completes with
+  `/v1`); glm's `unstructured` era re-attributed to the since-fixed v4.4.1 fence-extractor and
+  repair-prompt defects; every pre-degrade-era "silent shrink" claim re-grounded in the current
+  contract (one-voice announcement on stderr + `degrades[]` + `seatLoss`, exit 2, one Stage-1
+  retry with `Recovered:` heals). Upstreamed from the field: the peer-consensus≠evidence rule
+  (2 of 23 findings factually wrong inside high-trust tiers, caught only by primary-source
+  verification), the anti-narration-preamble and multi-file-stub rules, the Plan-hard-refusal
+  fallback, the stale-MCP-server signature, three model sections (qwen, gpt-5.6-terra,
+  gemini-3.1-pro), and gemini/deepseek/gpt/grok/kimi/Claude enrichments. Standing practice: each
+  release cherry-picks generalizable lessons per-section (docs/publishing.md release checklist),
+  never a bulk copy.

package/src/council/run-launch.js

@@ -94,6 +94,10 @@ function createLaunchers(deps = {}) {
       // part of that allowance in the meantime. This is the CLAIM that settles
       // it — synchronous by contract, so two callers can never interleave.
       ...(reserveBudget ? { reserveBudget: (est) => reserveBudget(opts.waveId, est) } : {}),
+      // SL-2: a Stage-1 retry names the wave it replaces; fanout threads this
+      // onto every leg and its spend-ledger row (v4.3 --retry-failed machinery).
+      // Spread-guarded so a normal launch's transport call stays byte-identical.
+      ...(opts.retryOfWaveId ? { retryOfWaveId: opts.retryOfWaveId } : {}),
       models: opts.models.join(','),
       prompt: opts.prompt,
       promptMeta: { source: 'council-engine', file: null, chars: opts.prompt.length },

package/src/council/run-retry-notes.js

@@ -0,0 +1,74 @@
+'use strict';
+
+/**
+ * @module council/run-retry-notes
+ * Pure note-builders for the SL-2 Stage-1 retry pass (split out of
+ * run-retry.js for the 300-line gate — same rationale as run-stage2.js
+ * splitting off run-stages.js, v4.4.1 Task 2). No I/O, no ctx: each function
+ * takes plain data and returns a still-dead note ready for
+ * `ctx.degrade.note(...)` (D5 final-failure granularity, spec §5). The heal
+ * note is built inline in run-retry.js's orchestrator (it is the one place
+ * that decides recovery, and stays small).
+ */
+
+/** D-effect parity: still-dead leg notes reuse today's count phrasing, with the
+ *  FIRST attempt's counts — the why carries the retry story (spec §5). */
+const legEffect = (counts) =>
+  `${counts.reviewed} of ${counts.total} seats reviewed; `
+  + 'the run continues with the bench that did and will exit degraded (2)';
+
+/** Wave-origin, retry wave died wholesale (D5 wave granularity). */
+function waveStillDeadNote(w, unit) {
+  return { channel: 'dead-wave',
+    what: `Stage-1 wave ${w.waveId} (${(w.models || []).join(', ') || 'no models'}) produced NO legs`,
+    // Coordinator-review MINOR-7c: a falsy w.reason must not render as the
+    // literal string "undefined" in the why text.
+    why: `${w.reason || 'no reason recorded'}; the once-only retry wave also produced no legs`,
+    effect: 'Those seats are NOT in this council. The run continues with the bench that did '
+      + 'launch and will exit degraded (2)',
+    data: { waveId: w.waveId, models: w.models, reason: w.reason, retryWaveId: unit.waveId } };
+}
+
+/** Leg-origin, retry wave died wholesale (bench-batch case). */
+function srcLegStillDeadNote(leg, unit, counts) {
+  const seat = leg.modelInput || leg.model;
+  return { channel: 'dead-leg', what: `seat ${seat} did not review`,
+    why: `the leg ended '${leg.status}'${leg.error ? `: ${leg.error}` : ''} with no usable output; `
+      + 'its once-only retry wave produced no legs',
+    effect: legEffect(counts),
+    data: { seat, status: leg.status, reason: leg.error || null, retryWaveId: unit.waveId } };
+}
+
+/** Either origin, the retry produced legs but THIS seat's retry leg died. */
+function retryLegStillDeadNote(seat, ff, retryLeg, unit, counts) {
+  const why = ff && ff.class === 'wave'
+    ? `its first wave ${ff.waveId} produced no legs (${ff.reason}); `
+      + `its once-only retry leg ended '${retryLeg.status}' with no usable output`
+    : `the leg ended '${ff ? ff.status : 'unknown'}'${ff && ff.reason ? `: ${ff.reason}` : ''} `
+      + `with no usable output; its once-only retry also ended '${retryLeg.status}'`;
+  return { channel: 'dead-leg', what: `seat ${seat} did not review`, why,
+    effect: legEffect(counts),
+    data: { seat, status: retryLeg.status, reason: retryLeg.error || null,
+      firstFailure: ff, retryWaveId: unit.waveId } };
+}
+
+/**
+ * CRITICAL fix (coordinator review): a launched seat can be missing a leg
+ * record ENTIRELY from the retry response — a partial wave return (unit
+ * models [a,b], the wave comes back with only a's leg). This is distinct
+ * from `retryLegStillDeadNote` (the seat's retry leg came back but was
+ * unusable) — here there is no retry-attempt status/error to report at all,
+ * only the ORIGINAL first-failure fact plus the fact that nothing came back
+ * this time.
+ */
+function missingLegStillDeadNote(seat, ff, unit, counts) {
+  const fact = ff && ff.class === 'wave'
+    ? `its first wave ${ff.waveId} produced no legs (${ff.reason})`
+    : `the leg ended '${ff ? ff.status : 'unknown'}'${ff && ff.reason ? `: ${ff.reason}` : ''} with no usable output`;
+  return { channel: 'dead-leg', what: `seat ${seat} did not review`,
+    why: `${fact}; its once-only retry produced no leg for this seat`,
+    effect: legEffect(counts),
+    data: { seat, status: null, reason: null, firstFailure: ff, retryWaveId: unit.waveId } };
+}
+
+module.exports = { waveStillDeadNote, srcLegStillDeadNote, retryLegStillDeadNote, missingLegStillDeadNote };