@event4u/agent-config 16.3.0 → 16.5.0

package/dist/agent-src/commands/review/changes.md

@@ -39,10 +39,14 @@ to make sure the current branch is up to date with its base chain:
   - If exactly one open PR is found → invoke `/prepare-for-review`
     with that PR number. It will update `main`, fetch and merge the
     full branch chain into the current branch, and leave the current
-    branch checked out.
+    branch checked out — under `git.update_strategy: rebase` it merges
+    nothing and reports the behind branches instead.
   - If no open PR is found → fall back to a minimal local update:
     `git fetch origin main` and `git merge origin/main --no-edit` on
-    the current branch. Abort on conflict and report.
+    the current branch. Abort on conflict and report. Under
+    `git.update_strategy: rebase` skip the merge: `git fetch origin main`
+    only, review the diff against `git merge-base HEAD origin/main`, and say
+    the branch is behind.
   - If multiple PRs are found → ask the user which PR to use before
     proceeding.
 - If `/prepare-for-review` aborts (merge conflict, network error,
@@ -324,23 +328,12 @@ before re-review, hand off to
 generate→run→revise→re-run loop gated by a numeric threshold, then a judge
 confirms. Surface it as a single follow-up line; do not auto-invoke.
 
-### 7. Quality tools (verbosity-gated)
+### 7. Quality tools
 
-Per `verbosity.routine_confirmations` (default `false`):
-
-- `false` (default) → emit `→ run /quality-fix to format + lint` as a
-  single follow-up line; do not auto-invoke. User runs explicitly.
-- `true` → ask:
-  ```
-  > 1. Yes — run quality tools (formatter, static analyzer, linters)
-  > 2. No — review done
-  ```
-  If yes, hand off to the project's quality workflow (e.g.
-  `/quality-fix` or the equivalent configured command). The quality and
-  test runners are resolved per-stack via the
-  [`toolchain-resolver`](../../contexts/execution/toolchain-resolver.md), so
-  the hand-off adapts to PHP / JS-TS / Python / Go / Rust rather than
-  assuming one stack.
+Emit `→ run /quality-fix to format + lint` as a single follow-up line; do not
+auto-invoke and do not ask. The user runs it explicitly; its quality and test
+runners are resolved per-stack via the
+[`toolchain-resolver`](../../contexts/execution/toolchain-resolver.md).
 
 ## Backward compatibility
 
@@ -354,7 +347,9 @@ Per `verbosity.routine_confirmations` (default `false`):
   quality tools hand-off
 - The new step 1 (`/prepare-for-review`) is **best-effort**: if no
   open PR exists for the current branch, it falls back to a plain
-  `git fetch && git merge origin/main`. Existing invocations that
+  `git fetch origin main` and, under the default `git.update_strategy:
+  merge`, `git merge origin/main --no-edit`; under `rebase` it merges
+  nothing and reports the branch as behind. Existing invocations that
   ran on a fully detached or pre-PR branch keep working
 
 ## Use this command when

package/dist/agent-src/commands/roadmap/next.md

@@ -280,7 +280,16 @@ was for a conflict whose entire existence is a remote fact. That is the friction
 this clause removes; the floor it keeps is the *unnamed* push.
 
 **Before every push to the open PR**, bring the branch up to its base with
-`./scripts-run src/scripts/sync_pr_branch` and regenerate afterwards. Measured on
+`agent-config git:convention sync --base origin/<base>` (`<base>` is the
+`baseRefName` of the branch's `OPEN` pull request, read with `gh pr view --json
+state,baseRefName`; without an open one, and without `--base`, the target is the default branch) and regenerate afterwards (under
+`git.update_strategy: rebase` it refuses a behind branch with exit 3 — rebase on request instead, per
+[`branch-update`](../../../../skills/git-workflow/references/branch-update.md); an exit 3
+reading `TARGET_POLICY_STALE` means the branch is current with its non-default
+target and the target is itself behind its default branch — that is the target's update, reported, never a rebase of this
+branch; on exit 4 the
+strategy could not be read, the line names the reason code and the file, nothing
+was merged, and the run stops before the push until that file is fixed). Measured on
 this run: the base moved three times, the push was rejected twice for it, and the
 PR reached `CONFLICTING` in between. Verify the result afterwards with
 `./scripts-run src/scripts/check_pr_ci_current` — a green local gate says nothing

package/dist/agent-src/commands/roadmap/process-full.md

@@ -7,7 +7,7 @@ cluster: roadmap
 sub: process-full
 skills: [agent-docs-writing, ai-council, roadmap-management]
 description: Autonomously process every open step across every phase of a roadmap until the file is fully closed. Largest execution scope of the /roadmap cluster — runs continuously across phase boundaries.
-argument-hint: "[roadmap] [--all] [--worktree]"
+argument-hint: "[roadmap] [--all] [--merge] [--worktree]"
 suggestion:
   eligible: false
   rationale: "Cluster sub-command — reached via its cluster head's routing or its explicit /cluster:sub name; not independently suggested (surface-consolidation)."
@@ -149,9 +149,9 @@ On outcome `complete`, open the PR as today, then run
 base in, resolve conflicts by that command's four enumerated classes, drive the
 required checks green on the pushed head. **This is unconditional** — a bare
 `/roadmap:process-full` delivers a mergeable PR, not merely an open one, and
-waiting on that remote CI is part of the run per the Iron Law below. It is
-the whole of what this command delivers: there is no flag that carries it
-further, and § Merging below says why.
+waiting on that remote CI is part of the run per the Iron Law below. With an
+owner auto-merge instruction the call is `/pr:merge <N> --merge` instead, which
+merges through that command's danger gate — § Merging below.
 
 ### The six required layers, in order, each with the command that runs it
 
@@ -193,39 +193,31 @@ condition to test rather than assert: a PR whose required context is skipped has
 a green rollup and an unenforced gate, so a delivery-ready test that reads only
 the rollup passes exactly when it should fail.
 
-### Merging — out of scope, cancelled rather than deferred
+### `--merge` — merging, on the owner's word only (ADR-282)
 
 ```
-THIS COMMAND NEVER MERGES. THERE IS NO FLAG THAT MAKES IT MERGE.
-THE RUN ENDS AT MERGEABLE-AND-OPEN, ALWAYS, AND SAYS SO.
-A FUTURE MERGE CAPABILITY NEEDS A NEW ROADMAP AND AN OWNER RULING —
-NOT A COMMAND EDIT, AND NOT A FLAG THAT ALREADY SITS HERE INERT.
+THIS COMMAND MERGES ONLY ON AN EXPLICIT OWNER AUTO-MERGE INSTRUCTION:
+`--merge`, OR THE OWNER'S OWN WORDS IN THE SESSION THAT RUNS THE WORK.
+WITHOUT ONE, THE RUN ENDS AT MERGEABLE-AND-OPEN, AND SAYS SO.
+A `needs-council` PR MERGES ONLY ON TWO CONVERGENT CLEARING SEATS;
+DANGEROUS, SPLIT, DEGRADED OR UNAVAILABLE → THE OWNER DECIDES.
 ```
 
-A `--merge` flag was specified here and has been **removed**. The
-`merge-authority` blocker of `road-to-drain-commands` closed as *not authorized
-in this roadmap*: activating the flag would have lowered
-[`non-destructive-by-default`](../../../../rules/non-destructive-by-default.md)'s
-per-turn confirmation floor for a production-branch merge, which
-[`decision-revisit-gate`](../../../../rules/decision-revisit-gate.md)'s
-owner-reserved table reserves to the owner. No owner ruling was available to the
-run that closed the roadmap, and an AI-council pass (2/2 convergent, 2026-08-22)
-ruled that a council may cancel an unauthorized implementation but may not
-manufacture the owner decision. The policy question is therefore **undecided,
-not rejected** — see
+The owner settled the half of
 [ADR-239](../../../../../docs/decisions/ADR-239-drain-command-surface-and-merge-authority.md)
-§ Disposition.
-
-The flag was removed rather than left inert deliberately: an archived roadmap
-must not leave latent executable authority behind a documented switch.
-
-Three independent findings stand behind the safety requirement, and any future
-attempt starts from them rather than re-deriving them: the AI council's Q1
-verdict (2026-08-21 — mergeability-only until authorization is target-bound and
-tamper-resistant), the committed `road-to-gate-preauth-authorization` stub (an
-authorization the agent can write is not an authorization), and the runtime
-guard that refused this roadmap's own attempt to edit the canonical loop
-contract.
+§ Disposition that ADR-266 left open:
+[ADR-282](../../../../../docs/decisions/ADR-282-auto-merge-authority-with-a-council-danger-gate.md).
+An explicit auto-merge instruction is the this-turn confirmation
+[`non-destructive-by-default`](../../../../rules/non-destructive-by-default.md)
+requires for the production-branch merge of a **routine** PR. Delivery runs
+`/pr:merge <N> --merge`, whose § 9a danger gate (`classify_merge_risk`) decides:
+`routine` → squash-merge bound to the observed head; `needs-council` → a neutral
+council review, merge only when both seats say *not dangerous, owner review
+unnecessary*, otherwise hand the PR back with the seats' reasons.
+
+ADR-268 § 3's object-bound `prod_merge` grant stays a separate, unchanged path.
+A standing autonomy directive or momentum is **neither**. Deploy, release and
+every other Hard-Floor action keep their own confirmation.
 
 **Mergeability is per-PR against a recorded base, never a queue property.**
 When every PR in the estate touches the same generated files — in this
@@ -481,7 +473,12 @@ A missing setting, a CI re-run, a protection flag, a rebase, a council call: do
 them, and say so in the PR body if it matters. The only mid-run interruptions
 that remain legitimate are a spend-ceiling crossing and an EXCLUDED-list action
 (production trunk, deploy, prod data/infra, an irreversible external action
-beyond the PR).
+beyond the PR). One carve-out from "do them": a rebase of a branch that is
+already pushed rewrites published history, and still needs the user's answer
+this turn per `git-history-discipline` § When rewrite is allowed — the
+invocation does not give it, and neither does `git.update_strategy: rebase`,
+which picks the operation and never authorises it. Without that answer the
+branch is left behind and the PR body says so.
 
 Report shape:
 
@@ -500,18 +497,35 @@ Hard external blocker: <specific reason>
 Attempted: <what was tried>
 ```
 
-## Spend — USD 25 per run, pre-authorised
+## Spend — whatever ceiling was configured, and none when none was
 
 ```
-cost_so_far + reasonably_expected_remaining  ≤  $25  → RUN IT, ASK NOTHING
-cost_so_far + reasonably_expected_remaining  >  $25  → OWNER APPROVAL BEFORE CROSSING
+READ THE CONFIGURED cost.budgets WINDOWS (daily · weekly · monthly).
+spend_so_far + reasonably_expected_remaining  WITHIN EVERY CONFIGURED WINDOW
+    → RUN IT, ASK NOTHING
+WOULD CROSS A CONFIGURED WINDOW → PAUSE AND REPORT BEFORE CROSSING.
+NO WINDOW CONFIGURED → NONE APPLIES. RUN IT, ASK NOTHING.
 COST UNKNOWN → CONSERVATIVE ESTIMATE. UNCERTAINTY IS NOT A REASON TO ASK.
-SERVICE-ENFORCED LIMITS MAKE >$25 IMPOSSIBLE → AUTHORISED.
+SERVICE-ENFORCED LIMITS MAKE A CROSSING IMPOSSIBLE → AUTHORISED.
 EXISTING SUBSCRIPTION → MARGINAL COST $0, NOT COUNTED.
 NEVER SPLIT SPEND ACROSS SERVICES, SUBAGENTS, COUNCIL ROUNDS OR CALLS
-TO KEEP EACH ITEM UNDER THE CEILING. THE CEILING IS CUMULATIVE, PER RUN.
+TO KEEP EACH ITEM UNDER A CEILING. A CEILING IS CUMULATIVE, PER RUN.
 ```
 
+**Who carries this, stated plainly: the model, not a gate.** Nothing on a
+roadmap run reads a cumulative spend figure and compares it with anything —
+ADR-237 said so of its own ceiling ("Nothing mechanical checks any of this")
+and that has not changed. The optional `cost.enforcement: hard-stop` preflight
+is a separate, deliberately-invoked control and is not called from a run. So
+this block is an obligation the model carries, and a reader should not take it
+for an enforced limit.
+
+What used to sit here was a flat per-run USD constant that nobody configured.
+ADR-279 replaced it with the windows a person actually set, which is the only
+kind of ceiling this package now applies; that record names the old figure, and
+this file deliberately does not, so no reader mistakes a historical number for
+a live threshold. Set `cost.budgets.daily` / `weekly` / `monthly` to get one.
+
 The transport is not a governance boundary: a council invocation is the same
 authorised action whether it runs through the `council` CLI or a direct API call.
 There is deliberately **no allowlist** of blessed services — the general rule

package/dist/agent-src/commands/roadmap/resolve-blockers.md

@@ -36,6 +36,10 @@ AS A CHECK BEFORE TREATING IT AS OPEN.
 RESOLVE AT THE LOWEST RUNG THAT OWNS THE DECISION: EVIDENCE → RECORDED RULE →
 EXECUTION → AGENT → INDEPENDENT SESSION → COUNCIL → OWNER.
 THE OWNER IS THE LAST RUNG, NEVER THE FIRST, AND NEVER FOR A TECHNICAL DECISION.
+NO OWNER QUESTION IS PUT BEFORE ONE BATCHED COUNCIL PASS HAS RUN OVER EVERY
+UNDECIDED BLOCKER OF THE RUN — OWNER-LABELLED ONES INCLUDED. A BLOCKER'S
+`Owner:`, `Ownership:` OR "YOURS TO SET" IS ITS AUTHOR'S GUESS; THE COUNCIL'S
+OWNER-RESERVED TEST DECIDES WHO MUST ANSWER.
 NEVER PASS `--confirm` OR BUY METERED COUNCIL SPEND WITHOUT THE OWNER'S YES.
 THE OWNER IS ASKED ONE QUESTION PER TURN, WITH OPTIONS AND ONE RECOMMENDATION,
 AND THE ANSWER IS WRITTEN DOWN BEFORE THE NEXT QUESTION IS PUT.
@@ -87,7 +91,10 @@ at all. Re-classifying a technical `3` to `2` is a legitimate outcome of this
 run, recorded on the entry (`Class: 2 — council-decidable. Corrected from 3 on
 YYYY-MM-DD`), written as the bold field `- **Class:** 2 — …` the lint reads.
 Read each entry's `- **Ownership:**` too: `product-owned`, `business-owned` and
-`destructive-owned` are locked to the owner whatever the class says.
+`destructive-owned` are the author's routing claim, written when nobody had run
+the test. They mark a candidate for the owner, never a destination: the
+blocker still goes into the council batch, whose owner-reserved test (§ Running
+the council) decides whether it reaches the owner.
 
 `unblocksSteps` is the open-step count of the whole roadmap the blocker sits in,
 the same number for every blocker of that roadmap. Order by the steps a blocker
@@ -151,11 +158,14 @@ uses, with one blocker-specific rung added (execution):
    (`decision-revisit-gate` sends those council-first). See § Running the
    council. Where no council is available, the **team** rung (`ai_team`) takes
    its place.
-7. **The owner** — **only** `product-owned`, `business-owned` and
-   `destructive-owned` decisions, and the transitions `decision-revisit-gate`
-   reserves (lowering a safety floor, an irreversible or external commitment,
-   spend above a delegated threshold). A technical decision never reaches here
-   by being hard. One exception, from the ownership contract: a
+7. **The owner** — **only** what the council's batch pass marked
+   owner-reserved (either seat), or split on with the split surviving a
+   measurement: `product-owned`, `business-owned` and `destructive-owned`
+   decisions, and the transitions `decision-revisit-gate` reserves (lowering a
+   safety floor, an irreversible or external commitment, spend above a
+   delegated threshold, governance self-amendment). A technical decision never
+   reaches here by being hard, and an owner-labelled one never reaches here
+   without the council's pass. One exception, from the ownership contract: a
    `critical-technical` question with a single provider present degrades to
    **owner-confirm of the agent's proposal**, never to one model reviewing
    itself. See § Asking the owner.
@@ -182,10 +192,31 @@ Council availability is the resolver's answer, never the project tree's:
 agent-config council:status
 ```
 
-1. **Batch the questions.** All council-owned blockers of a run go into **one**
-   question file under `agents/runtime/council/questions/` — one numbered
+1. **Batch the questions — all of them, before any owner question.** Every
+   blocker rungs 1–4 did not close goes into **one** question file under
+   `agents/runtime/council/questions/`, including those labelled `Owner: user`,
+   `product-owned`, `business-owned` or `destructive-owned` — one numbered
    section per blocker, each carrying the blocker's five fields, the Phase-2
-   evidence and the candidate options. One run, not one per blocker.
+   evidence and the candidate options. One run, not one per blocker. The file
+   opens with the owner-reserved test, quoted from `decision-revisit-gate`'s
+   owner-reserved table: the chosen option would change the project's purpose
+   or a declared non-negotiable outcome; **lower** or remove a recorded
+   security / privacy / safety / data-handling floor; be irreversible or
+   materially destructive; spend or carry liability above a delegated
+   threshold; create, remove or weaken a legal, contractual, licensing,
+   compatibility or public commitment; or amend governance itself (reopening
+   authority, quorum, escalation, the owner-reserved set). Strengthening a
+   floor, or anything reversible inside these bounds, is council-decidable.
+   Each section then asks for three answers, in this order:
+   1. **Owner-reserved?** — yes / no, naming the clause, per option where the
+      options differ (one option can be reserved while another is not);
+   2. **Choice** — the option the seat takes if it is council-decidable, and
+      the option it would recommend to the owner if it is not;
+   3. **The fact that would change the answer.**
+   A recorded decision that reserved the question to the owner (a `## Decisions`
+   row, an ADR) is put to the same test, as a reopening per
+   `decision-revisit-gate` — the council says whether the reservation still
+   holds; it never lifts one on its own authority.
 2. **Write the question neutrally** — scope, evidence, options. No
    recommendation and no expected verdict, in either direction: the council is
    judging a decision the agent would otherwise take, so the agent's lean stays
@@ -205,7 +236,13 @@ agent-config council:status
    checkout that has run a council first — without it every seat reads as
    unavailable. `--output` must sit under `agents/runtime/council/responses/`;
    any other directory is refused. A failed run still spends quota, so never
-   re-run an unchanged question.
+   re-run an unchanged question — with one exception: a seat that is absent
+   for a **transport** reason (the response's `absent_members[].detail` names
+   an OS, network or process error, not a refusal or a verdict) allows **one**
+   transport retry of the unchanged file under a new `--output` name. The
+   retry is recorded as such, and if a seat that answered both times changed
+   its answer, both readings are reported. A second retry, or a retry after an
+   unwelcome verdict, is verdict shopping.
 
    **Exhausted quota is `spend-exhaustion`: pause and report** — which questions
    needed the council, the quota reading, `estimate --mode-override api`'s cost,
@@ -215,7 +252,10 @@ agent-config council:status
    owner has said yes to the metered rung, `agent-config council:grant-billing
    <run_id>` records that yes for one run, and only then does
    `--mode-override api` run.
-4. **Read the verdict honestly.** 2/2 agreeing → decided. 1/2 present is a
+4. **Read the verdict honestly.** Route first: a blocker both seats call **not**
+   owner-reserved is decided by the council's choice and never reaches the
+   owner; one that either seat calls owner-reserved goes to the owner, carrying
+   the seats' recommended option. Then the choice: 2/2 agreeing → decided. 1/2 present is a
    single-seat reading (`DEGRADED`), never a convergence: accept it for a
    `reversible-technical` or `contested-technical` question and say so; for a
    `critical-technical` one it degrades to owner-confirm of the proposal.
@@ -239,12 +279,18 @@ EVERY QUESTION IS ANSWERABLE WITH ONE NUMBER.
 THE ANSWER IS RECORDED AND APPLIED BEFORE THE NEXT QUESTION IS PUT.
 ```
 
+The queue holds only what the council's batch pass routed here (§ Running the
+council, step 4) — never a blocker the pass did not see, and never one both
+seats called council-decidable. With `--no-council`, or when no seat is
+available, say so in the queue line: the routing is then the agent's, unchecked.
+
 Before the first question, one short line says how many owner questions the run
 has left after the council (`3 decisions need you — here is the first`), so the
 owner knows the length of the queue. Then, per question:
 
 - **Context in two sentences** — what is blocked and what it unblocks.
-- **What the council said**, if it ran: verdict or the split.
+- **What the council said** — why it is owner-reserved (the clause each seat
+  named), and the option each seat would recommend, or the split.
 - **What exactly you do**, if the answer needs an act only the owner can
   perform — a repository setting, a ruleset toggle, a credential, a tag: the
   literal step (the page, the field, the command), so nothing has to be looked
@@ -345,11 +391,15 @@ council's to answer. The ledger and the inline council record are the control.
 - Mark a blocker resolved because its status is old, or because a sibling PR
   looks related. The condition, executed, is the only evidence.
 - Ask the owner what the tree, a recorded decision or the council can answer —
-  *"Owner: user"* in the file is not a reason.
+  *"Owner: user"*, `Ownership: product-owned` or *"yours to set"* in the file
+  is not a reason.
+- Put the first owner question before the council's batch pass has run over
+  every undecided blocker, or classify a blocker as owner-reserved yourself
+  when a council seat is available to run the test.
 - Ask more than one question per turn, or hand back *"four decisions are open in
   file X"*.
 - Run the council once per blocker, or re-run an unchanged question after a
-  failure.
+  failure — beyond the single transport retry § Running the council allows.
 - Pass `--confirm` or switch to metered council spend without the owner's yes.
 - Cite `agents/runtime/council/` from a roadmap.
 - Execute the work a resolved blocker unblocks — this command authorizes

package/dist/agent-src/commands/worktree/create.md

@@ -39,9 +39,13 @@ Without a primitive, the skill's plain `git worktree add` path applies.
 
 ### 2. Branch naming — propose once
 
-Derive ONE branch name from the task (`<type>/<short-slug>` per the
-`commit-conventions` rule), state it, and proceed. Do not present a
-naming menu; the user can override by naming a branch in the same turn.
+Derive ONE type, slug and (if any) ticket from the task and render the name
+through `git.branch_pattern`:
+`agent-config git:convention branch --type <type> --slug <slug> [--ticket <ID>]`
+prints the name; exit `1` names the value or pattern it refused — fix the input,
+never hand-render ([`commit-subject`](../../../skills/git-workflow/references/commit-subject.md) § Rendering a branch name). State the name and proceed. It goes into every command
+double-quoted (`-b "<branch>"`). Do not present a naming menu; the user can
+override by naming a branch in the same turn.
 
 ### 3. Write the scope-lock note
 

package/dist/agent-src/commands/worktree/verify.md

@@ -66,4 +66,5 @@ whether checklist item 3 (evidence attached) now passes.
 
 - **Do NOT commit or push.** Evidence lives in the untracked scope note.
 - Red probe → report it plainly and stop; never downgrade to a weaker
-  probe to get green (N=3 budget per `autonomous-execution`).
+  probe to get green (the `execution.fix_loop_max` budget and its ladder,
+  per `autonomous-execution`).

package/dist/agent-src/contexts/communication/rules-auto/think-before-action-mechanics.md

@@ -82,9 +82,11 @@ Never load full datasets into context. Extract what you need:
 - Do NOT loop through trial-and-error when one targeted inspection
   would reveal the cause.
 - Max 2 retries for the same approach — then stop and rethink. The 2
-  retries plus the initial attempt ARE the three attempts of
-  `autonomous-execution`'s N=3 stop: the 3rd consecutive failure on the
-  same target escalates to stop-and-ask, then to `context-hygiene`'s
+  retries plus the initial attempt ARE the first three attempts of
+  `autonomous-execution`'s `execution.fix_loop_max` ladder: the 3rd
+  consecutive failure on the same target closes band 1 and triggers the
+  rule's mandatory strategy shift — never a halt, a count alone is
+  not a reason to ask — then, after repeated EPOCHS, to `context-hygiene`'s
   fresh-session recommendation — one ladder, not competing budgets
   ([`autonomy-mechanics § Retry-budget escalation ladder`](../../execution/autonomy-mechanics.md)).
 

package/dist/agent-src/contexts/communication/rules-auto/token-efficiency-mechanics.md

@@ -54,23 +54,21 @@ command says "ask the user", you ask.
 - Max 3 diagnostic commands per error. Read the error, think, act.
 - One hypothesis at a time. Pick the most likely, try it. Fails →
   next hypothesis; failures on the same target accrue to the ONE
-  retry ladder (2 retries per approach → N=3 stop-and-ask, per
+  retry ladder (2 retries per approach → the 3rd closes band 1 and
+  triggers a mandatory strategy shift, never a halt, per
   [`autonomy-mechanics § Retry-budget escalation ladder`](../../execution/autonomy-mechanics.md))
   — never a separate one-strike cap.
 
 ### Keep intermediate output minimal
 
-Read `personal.minimal_output` (default: `true`) and
-`personal.play_by_play` (default: `false`) from `.agent-settings.yml`.
-
-When `personal.minimal_output: true`:
+Not a setting — this is the behavior
+[`direct-answers`](../../../rules/direct-answers.md) Iron Law 3 already
+mandates:
 
 - Multi-step work: short bullet points only, no paragraphs.
 - No thinking out loud — user doesn't need your reasoning.
-- `personal.play_by_play: false` → silently investigate, report
-  conclusion only.
-- `personal.play_by_play: true` → briefly share intermediate
-  findings.
+- Silently investigate, report the conclusion only. Share intermediate
+  findings only when the user asks for them in the turn.
 - At the end: concise summary — what changed, what user needs to
   know.
 

package/dist/agent-src/contexts/contracts/command-suggestion-flow.md

@@ -36,7 +36,7 @@ match()    →  raw scored matches per eligible command
    │
    ▼
 rank()     →  apply floor, drop blocklisted, anti-noise heuristics,
-              cap at max_options, stable tie-break
+              cap at 4 options, stable tie-break
    │
    ▼
 apply_cooldown()  →  drop (command, evidence) pairs shown in window
@@ -123,11 +123,13 @@ track separately.
 
 | Scope | Key | Effect |
 |---|---|---|
-| Whole project | `commands.suggestion.enabled: false` | Layer fully silent; explicit slashes still work |
 | Specific command | `commands.suggestion.blocklist: [cmd1, cmd2]` | Those commands never appear; still callable |
 | Specific conversation | User types `/command-suggestion-off` | Disabled until user types `/command-suggestion-on` or chat ends |
 
-Per-command frontmatter overrides the global floor and cooldown:
+There is no project-wide off switch: `commands.suggestion.enabled` and the
+global floor / cooldown / option-cap keys were retired on 2026-10-09 with their
+defaults (on · `0.6` · 600 s · 4) as fixed behavior. Per-command frontmatter
+still overrides the floor and cooldown for that command:
 
 ```yaml
 suggestion:
@@ -195,8 +197,8 @@ Together they enforce:
 
 On Claude.ai Web / Skills API, the suggester package is **not** part
 of the standard bundle (T2 ZIPs ship rules + skills, not Python
-helpers). Treat `commands.suggestion.enabled` as `false` — degrade
-silently, never crash the turn. Local agents (Augment, Claude Code,
+helpers). Treat the layer as absent there — degrade silently, never crash
+the turn. Local agents (Augment, Claude Code,
 Cursor, Cline, Windsurf) get the engine via `scripts/`.
 
 ## See also
@@ -204,4 +206,4 @@ Cursor, Cline, Windsurf) get the engine via `scripts/`.
 - [`command-suggestion`](../../rules/command-suggestion-policy.md) — runtime rule
 - [`adr-command-suggestion.md`](adr-command-suggestion.md) — architectural decision
 - [`command-suggestion-eligibility.md`](command-suggestion-eligibility.md) — locked eligibility table
-- [`agent-settings`](../../templates/agent-settings.md) — `commands.suggestion.*` reference
+- [`agent-settings`](../../templates/agent-settings.md) — `commands.suggestion.blocklist` reference

package/dist/agent-src/contexts/contracts/frugality-charter.md

@@ -33,12 +33,15 @@ Live Phase 1 schema (see [`templates/agent-settings.md`](../../templates/agent-s
 
 | Key | One-line semantics |
 |---|---|
-| `verbosity.preview_artifacts` | When `false`, skip generated commit messages / PR titles / branch names preview before acting. |
-| `verbosity.routine_confirmations` | When `false`, skip "looks good?" gates on routine workflow steps. Iron-Law gates always ask. |
 | `verbosity.offer_council_in_delivery` | When `false`, delivery commands skip the optional "run Council on this?" offer. |
-| `verbosity.post_action_reports` | `off` / `minimal` / `full` — multi-line status blocks after a successful action. |
-| `verbosity.intent_announcements` | When `false`, writers must not generate "Let me…", "Now I will…", "Found it" openers. |
-| `telegraph.speak` | When `false` (the shipped default) the telegraph-speak rule does not ship at all; when `true` its own carve-out list states the grammar's scope — options blocks, Iron-Law literals, code, and paths stay byte-stable. |
+
+Five former hooks were retired on 2026-10-09 with their defaults as the fixed
+behavior, so writers state the behavior instead of citing a key: no preview of
+generated commit messages / PR titles / branch names before acting; no
+"looks good?" gate on a routine step with one obvious answer (Iron-Law gates
+always ask); after a successful action, a one-line confirmation, never a
+multi-line status block; no "Let me…", "Now I will…", "Found it" openers; and
+the telegraph-speak rule does not ship.
 
 Phase 10 adds `verbosity.script_output` (`silent` / `minimal` / `verbose`) and `verbosity.taskfile_command_echo` (`true` / `false`); writer artifacts cite this row when those land.
 

package/dist/agent-src/contexts/execution/auto-dispatch-classification.md

@@ -276,5 +276,5 @@ rung-3 degrade path, the recursive-dispatch guard and every ∅ case.
 
 - [`auto-orchestration-activation`](auto-orchestration-activation.md) — the enable/auto/manifest gate that runs before classification.
 - [`subagent-orchestration`](../../skills/subagent-orchestration/SKILL.md) — the modes this selects.
-- [`autonomous-execution`](../../rules/autonomous-execution.md) — the N=3 budget any LLM-classification v2 must respect.
+- [`autonomous-execution`](../../rules/autonomous-execution.md) — the fix-loop budget v2 must respect.
 - [`orchestration-telemetry`](orchestration-telemetry.md) § Registered always-on metrics — the judgment-ladder precision metric this ladder feeds.

package/dist/agent-src/contexts/execution/auto-orchestration-activation.md

@@ -17,7 +17,6 @@ set, and what the host can physically do.
 
 | Key | Default | Purpose |
 |---|---|---|
-| `subagents.downshift` | `true` | Route delegable sub-tasks to the lowest-capable tier (Phase 2). |
 | `subagents.quota_arbitrage` | `true` | Prefer a separate-quota-pool model where the host manifest allows it (Phase 2 bonus). |
 | `subagents.model_map` | `{}` | Optional per-tier model override for downshift routing. |
 | `emergency.orchestration_halt` | `false` | The ONE audited incident switch — see [`settings-classes.md`](../../docs/contracts/settings-classes.md) § "The one exception". NOT an activation gate: on-by-default, ceremony-free to arm, requires a non-empty `orchestration_halt_justification` to disarm. |

package/dist/agent-src/contexts/execution/evidence-discipline.md

@@ -164,8 +164,8 @@ explicitly and the default flips back only on a positive reuse signal).
 - **Redaction on write.** Promotion runs the `low-impact-corpus-privacy-floor`
   + `source-confidentiality` pattern set and **halts** on any hit — never
   silent-shares, never auto-rewrites.
-- **Promotion is hybrid.** A `public`/`vendor` card seen in ≥`auto_promote_threshold`
-  distinct repos triggers a one-tap **suggestion**; never a silent write.
+- **Promotion is hybrid.** A `public`/`vendor` card seen in ≥2 distinct repos
+  triggers a one-tap **suggestion**; never a silent write.
 
 ## Evidence v2 rollback target (Phase 0)
 

package/dist/agent-src/contexts/execution/non-interactive-contract.md

@@ -104,25 +104,22 @@ black box):
 
 ## 7. Kill-switch — disable auto-detection without a rollback
 
-Three override layers, narrowest wins:
+Two override layers, narrowest wins:
 
 1. **Per-invocation:** `--no-auto-detect` → fall back to the menu
    (interactive) or `missing_sub_command` (CI).
 2. **Per-orchestrator:** `auto_detect: false` in the command's
    front-matter → that orchestrator always shows the menu.
-3. **Global:** `commands.auto_detect: disabled | warn | enabled` in
-   `.agent-settings.yml` (default `enabled`; `warn` = detect but always
-   confirm).
 
-A misbehaving heuristic is disabled by config, not by reverting 6.1.0.
+There is no global switch: the `commands.auto_detect` settings key was retired
+on 2026-10-09 with its default (`enabled`) as the fixed behavior, so a
+misbehaving heuristic is disabled per orchestrator, not by reverting 6.1.0.
 
 ## 8. Rollback procedure (every orchestrator references this)
 
 1. Re-run with the explicit sub-command (`/judge solo`) — detection is
    skipped.
-2. Disable globally: `commands.auto_detect: disabled` in
-   `.agent-settings.yml`.
-3. Disable one orchestrator: set `auto_detect: false` in its
+2. Disable one orchestrator: set `auto_detect: false` in its
    front-matter — the change takes effect on the next projection/sync.
 
 ## See also

package/dist/agent-src/contexts/execution/rdp-gate.md

@@ -23,11 +23,11 @@ maintain a model list, and never ship heavy/light content variants (two variants
 
 ## Three signals (all table-free)
 
-1. **User settings** — read the `reasoning:` block in `.agent-settings.yml`:
+1. **User settings** — read `reasoning.enabled` in `.agent-settings.yml`:
    - `enabled: false` → the whole layer is inert. Stop here.
-   - per-component switch off (e.g. `components.verifier_default: false`) → that
-     component never fires.
-   - `auto_gate: false` → skip signal 3 (self-assessment); gate on signal 2 only.
+   - There is no per-component switch and no switch for signal 3: both were
+     retired on 2026-10-09 with their default (`true`) as the fixed behavior,
+     so every component follows signals 2 and 3 below.
 2. **Task signal** (knowable per turn, no model lookup):
    - **Skip** RDP when the task is trivial / short / fully-specified (rename,
      typo, format, one-line edit, list files, bump a version).