docks-kit 0.17.2 → 0.18.0

package/AGENTS.md

@@ -28,7 +28,8 @@ launcher can fall back to Bun source.
 
 | Path | Purpose |
 |------|---------|
-| `docks-kit` / `docks-kit.ps1` | POSIX and Windows CLI launchers. On supported hosts, each runs the matching binary in `cli/dist/` only when its `--version` matches `package.json`. Otherwise it runs Bun-from-source and auto-installs Bun plus `node_modules`. Hosts outside the support matrix fail before source fallback. The standalone platform release binary provides no-Bun recovery. |
+| `docks-kit` / `docks-kit.ps1` | POSIX and Windows CLI launchers. On supported hosts, each runs the matching binary in `cli/dist/` only when its `--version` matches `package.json`. Otherwise it runs Bun-from-source and auto-installs Bun plus `node_modules`, and the mismatch warning names the rebuild remedy. A launcher never deletes a binary: on a host without Bun it is the only remaining recovery path. Hosts outside the support matrix fail before source fallback. The standalone platform release binary provides no-Bun recovery. |
+| `cli/build-binaries.sh` | Compiles the six release binaries. It checksums the six known artifact names rather than globbing `docks-kit-*`, so a packed `docks-kit-<version>.tgz` is never attested as a release binary. It stamps `cli/dist/VERSION` only when the directory holds one version, and on the next build discards every artifact that stamp attributes to a different version, because a launcher refuses to run one and keeping it makes `SHA256SUMS` span two versions. An artifact with no stamp has unknown provenance and may be a hand-built or downloaded recovery binary, so the script keeps it and warns; `--prune` discards those too. Nothing here removes an artifact the stamp vouches for. |
 | `cli/src/engine-native/` | EngineNative implementation for `sync`, `model`, and `toolchain`; idempotent, flag-gated for destructive reconciliation |
 | `cli/src/engine-native/ompSync.ts` | omp file deployment, marketplace registration, and plugin synchronization |
 | `cli/src/commands/omp.ts` | `docks-kit omp` session launcher: renders the free-model run overlay and forwards args to omp |
@@ -54,7 +55,7 @@ launcher can fall back to Bun source.
 
 Codex SoT notes:
 - `SoT/.codex/AGENTS.md` deploys to `~/.codex/AGENTS.md` as global Codex instructions.
-- `SoT/.codex/config.toml` pins Codex to `model = "gpt-5.6-sol"`, sets normal and plan reasoning to `high` with concise summaries, and sets `model_verbosity = "low"`, `personality`, live top-level `web_search`, workspace-write sandboxing with sandboxed command network access, cross-session `memories` (+ dedicated note tools), `[agents]` subagent limits (`max_threads = 12`, `max_depth = 2` — intentionally above Codex defaults for broad parallel kit work; deeper recursion increases cost and predictability risk), a 128 KiB `project_doc_max_bytes` budget for the repo-side AGENTS.md chain (the global `~/.codex/AGENTS.md` is uncapped and not counted), and enables the two Docks plugins `docks@docks` and `plan-lifecycle@docks` (the shared plan lifecycle).
+- `SoT/.codex/config.toml` pins Codex to `model = "gpt-6-sol"`, sets normal and plan reasoning to `high` with concise summaries, and sets `model_verbosity = "low"`, `personality`, live top-level `web_search`, workspace-write sandboxing with sandboxed command network access, cross-session `memories` (+ dedicated note tools), `[agents]` subagent limits (`max_threads = 12`, `max_depth = 2` — intentionally above Codex defaults for broad parallel kit work; deeper recursion increases cost and predictability risk), a 128 KiB `project_doc_max_bytes` budget for the repo-side AGENTS.md chain (the global `~/.codex/AGENTS.md` is uncapped and not counted), and enables the two Docks plugins `docks@docks` and `plan-lifecycle@docks` (the shared plan lifecycle).
 - `SoT/.codex/rules/*.rules` deploys to `~/.codex/rules/` as kit-managed Codex command policy. This is Codex's equivalent of permission allow/prompt/block rules; user-learned approvals in `~/.codex/rules/default.rules` are preserved.
 - `SoT/.codex/plugins/marketplace.json` deploys to Codex's personal marketplace path at `~/.agents/plugins/marketplace.json`; when the `codex` CLI is available, sync reruns `codex plugin add <plugin@marketplace>` for enabled SoT plugins so stale cached installs are refreshed.
 - Codex `/import` can copy Claude hooks into `~/.codex/hooks.json`. `codexSync.ts removeRetiredImportedHooks, legacy SessionStart cleanup` removes only recognized hooks from retired docks-kit Claude settings, backs up a changed file, and preserves user-authored hooks. The current Claude SessionStart program emits the structured JSON shape shared by both tools.
@@ -63,18 +64,22 @@ Codex SoT notes:
 - The `codex` CLI binary is upstream-owned, not kit-owned. The official standalone installer keeps package metadata under `$CODEX_HOME/packages/standalone` and places the `codex` symlink in `~/.local/bin` by default; sync only warns with a download-then-run installer command when the CLI is missing. Existing installs can self-update with `codex update`; npm and Homebrew remain upstream alternatives.
 
 - Claude runtime settings are an authoring template with sentinels. `claudeRuntime.ts` materializes absolute Bun/script paths only after the shared `bun.ts` bootstrap is ready; `claudeSync.ts` writes all runtime assets before atomically committing settings, then prunes the legacy shell scripts and Stop hook. Native `rate_limits` is the sole quota source, so jq/curl/OAuth caches are not runtime dependencies. A missing Bun defers only this cutover and preserves legacy pointers/files.
-- Claude's deployed SoT defaults are `model: opus` and `effortLevel: high`; `advisorModel` is deliberately absent/off. `--claude-advisor=on` is the per-machine opt-in and writes `advisorModel: fable` after the settings merge. `opus` is the alias, not a pinned id: the `minimumVersion` floor of 2.1.219 ensures Claude Code can resolve it to the newest Opus its provider offers — Opus 5 on the Anthropic API or Opus 4.6 on Microsoft Foundry — instead of silently capping Anthropic API users at Opus 4.8 under the former 2.1.170 floor. Keeping the alias provides provider portability and tracks future Opus releases; the literal `claude-opus-5` is unavailable on Foundry. The floor also subsumes Fable 5's older 2.1.170 requirement.
+- Claude's deployed SoT defaults are `model: opus` and `effortLevel: high`; `advisorModel` is deliberately absent/off. `--claude-advisor=on` is the per-machine opt-in and writes `advisorModel: opus` after the settings merge. `opus` is the alias, not a pinned id: the `minimumVersion` floor of 2.1.280 makes Claude Code resolve it to Opus 5.5, the default Opus from that release, instead of capping at Opus 5 under the former 2.1.219 floor. Keeping the alias tracks the next Opus release without a kit change and stays portable across providers. The floor also subsumes Fable 5.1's 2.1.257 requirement.
 - Claude permission rules are pruned, not host-scoped. The kit deploys its `PowerShell(...)` deny and ask rules on every host, because Claude Code leaves the PowerShell tool opt-in off Windows rather than unavailable, and a host that enables it must already carry the guards. `claudeRetired.ts RETIRED_PERMISSION_RULES, exact retired-rule inventory` lists the rules the kit no longer ships. `claudeSync.ts syncRemovals, retired-permission pass` force-prunes those exact strings from the kit-managed `~/.claude/settings.json` on every sync, because `settings.ts mergeSettings, permission-array union` would otherwise keep a dropped rule forever. A different user-authored rule survives. To restore an exact retired rule for one checkout, put it in that checkout's `.claude/settings.local.json`. Claude Code resolves that file against the working directory and merges it over user settings. Sync never reads or writes the checkout-local file. Claude Code has no user-scope local settings file. For a machine-wide restoration, add the rule to `SoT/.claude/settings.json`. At the same time, remove the exact string from `claudeRetired.ts RETIRED_PERMISSION_RULES, exact retired-rule inventory`. `permissions.allow` holds four read-only entries (`Read`, `Glob`, `Grep`, and `WebSearch`) plus the working-directory edit rule `Edit(./)`. Broad shell allow rules are not restored, because an allow rule resolves before Claude Code's read-only command analyzer and the auto-mode classifier. `autoMemoryEnabled` and `autoDreamEnabled` are both `false`.
 
 omp SoT notes:
 - `SoT/.omp/AGENTS.md`, `config.yml`, `models.yml`, and `mcp.json` deploy to `~/.omp/agent/`.
 - `SoT/.omp/intercom.json` deploys to `$PI_CODING_AGENT_DIR/intercom/config.json`. The default root is `~/.pi/agent`.
 - `ompSync.ts syncMergedYaml` deep-merges `config.yml` through `ompYaml.ts mergeOmpConfig` and `models.yml` through `mergeOmpModels`. Both wrap one generic mapping merge; only the config wrapper prunes stale `retry.fallbackChains` wildcards.
+- `ompSync.ts` runs `ompRemovals.ts syncOmpRemovals, retired-key inventory` right after the config merge, and that pass force-prunes retired kit-owned keys from `~/.omp/agent/config.yml` on every sync, without `--reconcile`. The pass is required because `mergeOmpConfig` is additive, so removing a key from the SoT alone never removes it from a deployed file. A key retired with a recorded value is pruned only while the deployed value still matches that value, so a user edit survives; a key retired outright, such as `providers.webSearchOrder`, is pruned at any value.
 - `cycleOrder` ends with `astra` as its fifth stop. `modelRoles.astra` is `openai-codex/gpt-6-astra:xhigh`, and `modelTags.astra` is visible. Astra and Fable fall back to each other through concrete selectors. `modelRoles.fable` remains `anthropic/claude-fable-5-1:medium`, with visible `modelTags.fable`. The hidden `switch_fable` role uses the same Fable selector and keeps an empty fallback chain.
-- `modelRoles.task` is `openai-codex/gpt-5.6-sol:high`. The four reviewer entries in `task.agentModelOverrides` inherit it through `@task`. Only bundled `reviewer` and `security-reviewer` are discoverable OMP agents; `code-reviewer` and `plan-reviewer` stay dormant.
-- `SoT/.omp/models.yml` declares Astra's full `low, medium, high, xhigh, max` ladder with `defaultLevel: xhigh` as the worked provider ladder-override example. `ompYaml.ts mergeOmpModels` preserves deployed-only keys in `~/.omp/agent/models.yml`, because a user file may carry provider credentials. Whole-file replacement is wrong.
+- `modelRoles.task` is `openai-codex/gpt-6-sol:high`, and `advisor` is `openai-codex/gpt-6-sol:medium`. `smol`, `commit`, and `tiny` use `openai-codex/gpt-6-luna`. The four reviewer entries in `task.agentModelOverrides` inherit `task` through `@task`. Only bundled `reviewer` and `security-reviewer` are discoverable OMP agents; `code-reviewer` and `plan-reviewer` stay dormant. omp 18.2.9 does not list `gpt-6-sol` or `gpt-6-luna` in its `openai-codex` catalog, so it fuzzy-matches both selectors to the GPT-5.6 models without a warning until the catalog lists them. The owner deployed them anyway. `cli/docs/omp-models.md` records the checks.
+- The five Anthropic roles (`default`, `slow`, `plan`, `designer`, `vision`) and the five Anthropic retry chains use `anthropic/claude-opus-5-5` at the levels the role map records. `cli/docs/omp-models.md` carries the Artificial Analysis capture behind the role map, read 2026-09-22 at Intelligence Index v4.3.2 and Coding Agent Index v1.5 from the AA comparison-page metric tables. Opus 5.5 max has the highest index in that topic. AA has not measured speed or latency for Opus 5.5 max, GPT-6 Sol, or GPT-6 Luna.
+- `modelRoles.web` is `web/firecrawl` and `retry.fallbackChains.web` carries the explicit 20-entry provider order. The kit declares both keys because the legacy `providers.webSearchOrder` key is retired: omp expands it in memory into these two keys and then drops it, and never writes that expansion back to disk. An explicit chain replaces omp's built-in web order wholesale, so every entry left out is a provider omp never tries. The owner removed the seven entries that named older models (Gemini 2.5 Flash, Claude Haiku 4.5, GPT-5.6, GPT-5.5, and Grok 4.5); keep every other provider.
+- `SoT/.omp/models.yml` declares Astra's full `low, medium, high, xhigh, max` ladder with `defaultLevel: xhigh` as the worked provider ladder-override example. It also carries a temporary `anthropic.modelOverrides.claude-opus-5-5` block with that model's limits, ladder, and prices, because the shared catalog still serves the id as a stub with null limits and zero cost; remove the block once the catalog publishes the row. `ompYaml.ts mergeOmpModels` preserves deployed-only keys in `~/.omp/agent/models.yml`, because a user file may carry provider credentials. Whole-file replacement is wrong.
 - `SoT/.omp/AGENTS.md` carries the rule `Please remove all mannered prose.` Anthropic's Fable 5.1 prompting guide documents mannered prose as a Fable 5.1 behavior and gives that sentence as its short-version fix: https://platform.claude.com/docs/en/build-with-claude/prompt-engineering/prompting-claude-fable-5-1
 - `cli/docs/omp-models.md` (topic `omp-models`) records the role map rationale and the Artificial Analysis snapshot behind it. Model choices change with published benchmarks, so update that topic in the same commit as a role change.
+- `SoT/.omp/config.yml` sets `compaction.thresholdTokens` to `-1`, omp's schema default sentinel selecting reserve-based behavior: the trigger becomes `contextWindow` minus `max(floor(contextWindow * 0.15), 16384)`, which is 231,200 on the 272,000-token Codex window and 850,000 on the 1,000,000-token Anthropic window. The key ships as `-1` rather than being deleted because `ompYaml.ts mergeOmpConfig` is additive, so removing a key from SoT never removes it from a deployed file. `cli/docs/omp-context.md` (topic `omp-context`) carries the derivation and the measured evidence, so update that topic in the same commit as any compaction-setting change.
 - Sync registers the `docks` marketplace. It installs or upgrades `docks@docks` and `plan-lifecycle@docks` at user scope.
 - Sync installs `pi-intercom` at the verified version from `SoT/toolchain.json`.
 - The omp CLI is upstream-owned and self-updating through `omp update`. Sync never installs or upgrades the CLI.
@@ -115,7 +120,7 @@ For per-tool SoT layouts (`SoT/.claude/`, `SoT/.codex/`, `SoT/.omp/`), see the m
 - **One exemption: the kit's own package.** `install.sh` and `install.ps1` end with `bun add -g docks-kit@latest`, because a global installer that pinned itself would install a fixed old kit forever, and pinning it to `package.json` would request an unpublished version between the release-prep commit and the npm publish. The exemption covers `docks-kit` alone. Both installers still pin the Bun installer they download to the manifest's verified version, and `cli/test/unit/install.test.ts` asserts that pin in all four launcher and installer scripts.
 ## Testing
 
-Run `bun run check` before and after each change: oxlint, `tsc --noEmit`, unit tests, and both golden suites. Scripts: `lint`, `typecheck`, `test:unit`, `test:integration` (goldens), `test`, `format`; CI runs `check` on push and pull request with Bun 1.4.2. Prove-red modes must exit non-zero after detecting planted mismatches. Also verify user-facing changes via `./docks-kit sync --dry-run`, per-tool sanity (`/doctor`, `/plugin`, etc.), `bun run test:runtime:posix`, and `diff <(jq -S . <SoT>) <(jq -S . <deployed>)` recipes from the per-tool file.
+Run `bun run check` before and after each change: oxlint, `oxfmt --check`, `tsc --noEmit`, unit tests, and both golden suites. Scripts: `lint`, `format:check`, `typecheck`, `test:unit`, `test:integration` (goldens), `test`, `format`; CI runs `check` on push and pull request with Bun 1.4.2. Prove-red modes must exit non-zero after detecting planted mismatches. The golden writer emits two-space JSON that `oxfmt` re-serializes, so run `bun run format` after `--update-goldens` or `format:check` fails. Also verify user-facing changes via `./docks-kit sync --dry-run`, per-tool sanity (`/doctor`, `/plugin`, etc.), `bun run test:runtime:posix`, and `diff <(jq -S . <SoT>) <(jq -S . <deployed>)` recipes from the per-tool file.
 
 Use direct acceptance and focused regressions while iterating, then run the full unit/golden gate once at the pre-commit or release boundary. Reuse still-matching evidence; a later relevant edit invalidates only the affected rung and final gate, not every prior check.
 

package/cli/docs/models.md

@@ -17,15 +17,16 @@ the entry and date when a model ships or retires.
 
 ## The `best` alias and `default` pseudo-value
 
-- `best` resolves to Fable 5 where the org has access, latest Opus otherwise.
-  It needs Claude Code >= 2.1.170. The kit SoT pins the `opus` alias directly
-  rather than `best` or a full id, so the deployed model is unambiguous. Its
-  `minimumVersion` of 2.1.219 ensures Claude Code can resolve that alias to the
-  newest Opus its provider offers: Opus 5 on the Anthropic API or Opus 4.6 on
-  Microsoft Foundry. The former 2.1.170 floor silently capped Anthropic API
-  users at Opus 4.8. Keeping the alias provides provider portability and tracks
-  future Opus releases; the literal `claude-opus-5` is unavailable on Foundry.
-  The floor also subsumes `best`/Fable 5's older 2.1.170 requirement.
+- `best` resolves to Fable 5.1 where the org has access, latest Opus
+  otherwise. It needs Claude Code >= 2.1.257. Claude apps gateway sessions
+  still resolve `best` and `fable` to Fable 5. The kit SoT pins the `opus`
+  alias directly rather than `best` or a full id, so the deployed model is
+  unambiguous. Its `minimumVersion` of 2.1.280 ensures Claude Code can resolve
+  that alias to Opus 5.5, the default Opus from that release. The former
+  2.1.219 floor capped the alias at Opus 5. The kit keeps the alias because it
+  tracks the next Opus release without a kit change and stays portable across
+  every provider that carries a different newest Opus. The floor also subsumes
+  the older `best`/Fable 5.1 requirement of 2.1.257.
 - `default` is an engine pseudo-value: it DELETES the deployed `model` key so
   the account default applies. It never reaches the settings file as a value.
 
@@ -35,14 +36,13 @@ the entry and date when a model ships or retires.
 docks-kit models                  # both catalogs
 docks-kit models claude --json    # machine-readable
 docks-kit model claude            # current deployed + SoT + picker (TTY)
-docks-kit model claude opus       # per-machine override from the Fable SoT
+docks-kit model claude opus       # per-machine override from the Opus SoT
 docks-kit sync claude --claude-model=opus   # same, as part of a sync
 ```
 
 ## Advisor pairing note (Claude)
 
-The SoT ships `model: fable` with advisor off (`advisorModel` unset).
+The SoT ships `model: opus` with advisor off (`advisorModel` unset).
 Advisor is a per-machine opt-in: `docks-kit sync claude --claude-advisor=on`
-writes `advisorModel: fable`; `off` and `default` delete the key. Fable-main +
-Fable-advisor is an accepted pairing. The advisor needs Fable org access and
-Claude Code >= 2.1.170.
+writes `advisorModel: opus`; `off` and `default` delete the key. The advisor
+then runs the same Opus 5.5 the main session uses.

package/cli/docs/modifiers.md

@@ -9,15 +9,15 @@ the SoT is never touched. They all share one contract:
 > profiles can instead use `~/.claude/settings.local.json`, which sync never
 > touches.
 
-Claude's embedded SoT is `model: fable`, `effortLevel: high`, with advisor
+Claude's embedded SoT is `model: opus`, `effortLevel: high`, with advisor
 off (`advisorModel` unset). Codex's embedded normal and plan reasoning effort is
 `high`.
 
 | Modifier | Deployed change | Typical use |
 |----------|-----------------|-------------|
-| `--claude-model=<m>` | `.model` in ~/.claude/settings.json (`default` deletes the key) | Override one machine while the SoT retains `fable` |
+| `--claude-model=<m>` | `.model` in ~/.claude/settings.json (`default` deletes the key) | Override one machine while the SoT retains `opus` |
 | `--claude-effort=<level>` | `.effortLevel` in ~/.claude/settings.json (`default` writes `high`) | Tune persisted Claude effort per machine; valid `low`, `medium`, `high`, `xhigh` |
-| `--claude-advisor=<state>` | `on` sets `.advisorModel = "fable"`; `off`/`default` remove it | Enable Claude advisor only on machines that need it |
+| `--claude-advisor=<state>` | `on` sets `.advisorModel = "opus"`; `off`/`default` remove it | Enable Claude advisor only on machines that need it |
 | `--claude-compact-window=<n>` | `env.CLAUDE_CODE_AUTO_COMPACT_WINDOW` | Disposable containers running long autonomous work (e.g. `680k`) — not host machines |
 | `--claude-permissive` | `permissions.ask = []`, `permissions.deny = []` | Sandboxes/containers where prompts stall unattended work. Never on a host — the deny list is the safety floor |
 | `--codex-model=<m>` | top-level `model = "…"` in ~/.codex/config.toml | Same as claude-model, for Codex |
@@ -29,7 +29,7 @@ selecting that positional target warns and ignores it.
 
 A flag-less Claude sync also removes the formerly kit-owned `advisorModel`
 from machines synced before advisor became opt-in. Any explicit advisor state
-owns that key for the run: `on` writes `fable`; `off` and `default` delete it.
+owns that key for the run: `on` writes `opus`; `off` and `default` delete it.
 Codex has no advisor modifier because its documented config has no advisor
 setting; `review_model` applies only to `/review`.
 

package/cli/docs/omp-context.md

@@ -0,0 +1,104 @@
+# omp context and compaction
+
+This topic records why `SoT/.omp/config.yml` does not pin a compaction
+threshold, and what omp's reserve-based default computes on each model lane.
+
+## How omp resolves the trigger
+
+`resolveThresholdTokens` in omp source
+`packages/agent/src/compaction/compaction.ts` selects the trigger from three
+branches, in priority order.
+
+1. A positive `compaction.thresholdTokens` wins. omp clamps it to the range 1
+   through `contextWindow - 1`.
+2. Otherwise a positive `compaction.thresholdPercent` applies. omp clamps the
+   percent to the range 1 through 99 and computes
+   `floor(contextWindow * percent / 100)`.
+3. Otherwise, when both keys hold `-1`, reserve-based behavior applies.
+
+The reserve-based branch computes the trigger from the window itself:
+
+```text
+threshold = contextWindow - max(floor(contextWindow * 0.15), 16384)
+```
+
+omp names the 16384 constant `DEFAULT_RESERVE_TOKENS`. The 15% proportional
+reserve dominates on any window above roughly 109,227 tokens, so every lane
+the kit uses resolves through the proportional term and never through the flat
+constant.
+
+## Per-lane result
+
+| Lane | Window | Reserve | Threshold | Percent of window |
+|---|---:|---:|---:|---:|
+| `openai-codex/*` | 272,000 | 40,800 | 231,200 | 85.0% |
+| `anthropic/*` | 1,000,000 | 150,000 | 850,000 | 85.0% |
+
+Both lanes compact at the same fraction of their own window. The absolute
+trigger differs only because the windows differ.
+
+## Why the kit no longer pins 231200
+
+The kit previously shipped `compaction.thresholdTokens: 231200`. That value is
+exactly the reserve-based default for a 272,000-token window. It was therefore
+identical to the default on the Codex lane and changed nothing there. Its only
+real effect was forcing the Anthropic lane to compact at 23.1% of its window
+instead of 85%.
+
+The evidence below records 83 compactions measured between 2026-09-12 and
+2026-09-21, after omp 18.1.18 introduced the Anthropic server-side compaction
+lane.
+
+| Lane | Compactions | Maximum tokens before | Over window | Warnings |
+|---|---:|---:|---:|---:|
+| `anthropic` | 75 | 248,583 (24.9% of window) | 0 | 0 |
+| `openai-codex` | 8 | 263,591 (96.9% of window) | 0 | 0 |
+
+The median context freed was 64.8%. The pin was safe rather than dangerous.
+No session ran past its window, and omp raised no warning on either lane.
+Under the reserve-based default none of the 75 Anthropic compactions would
+have fired, because no session reached 850,000.
+
+## Why the key ships as -1 rather than being deleted
+
+`ompYaml.ts` `mergeOmpConfig` is additive. Its `mergeMappings, deployed-key
+retention loop` returns every deployed key that the SoT omits to the merged
+output, and the only drop predicate is `dropFallbackWildcard`, which is scoped
+to `retry.fallbackChains`.
+
+Deleting the key from the SoT would therefore leave the stale `231200` in
+every already-deployed `~/.omp/agent/config.yml` permanently. A present key
+holding the schema default sentinel `-1` overwrites that stale value on the
+next sync.
+
+Verify the deployed value after a sync:
+
+```bash
+omp config get compaction.thresholdTokens
+```
+
+The command must print `-1`.
+
+## The clamp hazard a fixed pin carries
+
+The fixed-token branch clamps the pinned value to `contextWindow - 1`. On any
+model whose window is below the pinned value, the reserve collapses to one
+token, so omp compacts only once the context is already full. The
+reserve-based default always leaves at least 15% of the window free.
+
+No current role model has a window below 272,000, so the hazard was latent
+rather than live. Not pinning removes it entirely, including for any future
+model with a smaller window.
+
+## Related settings left at their omp defaults
+
+| Setting | Value | Reason |
+|---|---|---|
+| `extendedContext` | `false` | omp 17.4.0 added the setting defaulting on, and a later release flipped the default to off. It caps a model carrying a premium long-context price tier at that model's standard-pricing window, so `openai-codex/gpt-5.6-sol` reports 272,000 rather than about 1,050,000. Enabling it makes a Codex window overrun structurally impossible. The cost is a cliff rather than a marginal rate: any request above the 272,000 threshold bills the whole request at input $10 instead of $4, cache read $1.00 instead of $0.40, and cache write $12.50 instead of $5.00. |
+| `compaction.idleEnabled` | `true` | omp default. Idle compaction fires below the reserve-based trigger on both lanes. A compaction whose reason is `idle` skips the progress guard. |
+| `compaction.idleThresholdTokens` | `200000` | omp default. |
+| `compaction.idleTimeoutSeconds` | `300` | omp default. |
+| `compaction.methodOrder` | `["remote","snapcompact","handoff","shake","soft"]` | omp default, remote first. Since omp 18.1.18 the Anthropic server-side lane handles Anthropic models, so snapcompact is a fallback rather than the primary path. |
+
+Every row records a decision to leave the omp default in place. None of these
+keys carries a kit value that differs from the default omp already applies.