@alexeiled/claude-router 0.7.1 → 0.8.0

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
@@ -1,7 +1,7 @@
1
1
  {
2
2
  "name": "router",
3
3
  "displayName": "jev-router",
4
- "version": "0.7.1",
4
+ "version": "0.8.0",
5
5
  "description": "Auto-picks the right Claude model for each turn — small models for quick edits, mid-tier for code, frontier for hard problems. Uses Jev to classify each request.",
6
6
  "author": { "name": "Alexei Ledenev", "url": "https://github.com/alexei-led" },
7
7
  "repository": "https://github.com/alexei-led/claude-router",
package/CHANGELOG.md CHANGED
@@ -1,5 +1,49 @@
1
1
  # Changelog
2
2
 
3
+ ## 0.8.0 (2026-10-01)
4
+
5
+ ### Changed
6
+
7
+ - **Sonnet 5.5 replaces Sonnet 5.** The `sonnet` model is now
8
+ `claude-sonnet-5-5`, at the same prices. A live probe through Claude Code
9
+ 2.1.286 shows it takes all five effort levels and all three request
10
+ features, so the gateway no longer adapts its history.
11
+ - **`medium` runs on Sonnet 5.5 at `xhigh`**, not on Opus 5.5 at `high`.
12
+ `/router:medium` changes with it. `high` stays Opus 5.5 at `xhigh`. From
13
+ 2026-09-23 to 2026-10-01, 15 of 16 escalations from Sonnet moved the
14
+ session onto Opus for good, at about $31.6 of cold cache writes; most were
15
+ triggered by environment or hook errors, not by the model. An escalation
16
+ from `low` now stays on Sonnet.
17
+
18
+ ### Upgrade
19
+
20
+ - If your `router.json` sets `models.sonnet.id` to `claude-sonnet-5`, also
21
+ set `"features": ["mid-conversation-system"]`. The new default lists all
22
+ three features, and Sonnet 5 answers the other two with 400.
23
+ - To keep the old `medium`, set
24
+ `"routes": { "medium": { "model": "opus", "effort": "high" } }` and change
25
+ `skills/medium/SKILL.md` to match.
26
+
27
+ ## 0.7.2 (2026-09-25)
28
+
29
+ ### Fixed
30
+
31
+ - `router.json` accepted a `policy.downgradePivotUsd` of 0 and a
32
+ `policy.downgradeSlope` that is not a number. A zero pivot made the
33
+ downgrade bar `NaN` whenever a switch cost nothing, and then no downgrade
34
+ ever passed. Both values now stop a new gateway from starting, as the
35
+ upgrade keys already did; `/router:status` names the field.
36
+
37
+ ### Changed
38
+
39
+ - The downgrade bar counts what a cheaper model saves, not only its cache
40
+ write. The tax used to count the input of the next request alone. It now
41
+ nets the output and cache-read savings of the next
42
+ `policy.downgradeHorizonTurns` turns (default 5). From a warm Opus at
43
+ `xhigh` to a cold Sonnet, with 104k of context and 4k of output, the bar
44
+ drops from 0.935 to 0.922 (1-hour TTL). It never goes below 0.90. A model
45
+ without an `output` price keeps the input-only tax.
46
+
3
47
  ## 0.7.1 (2026-09-24)
4
48
 
5
49
  ### Fixed
package/README.md CHANGED
@@ -54,9 +54,9 @@ flowchart LR
54
54
 
55
55
  A local gateway receives each request from Claude Code. It changes the model,
56
56
  the effort and the thinking setting. For Haiku, it also limits the output
57
- size. It does not change your prompt or your tools. Sonnet and Haiku do not
58
- accept all the request features that Claude Code uses for Opus, so the gateway
59
- adapts those parts of the history for them, as Claude Code itself does. Prompt
57
+ size. It does not change your prompt or your tools. Haiku does not accept all
58
+ the request features that Claude Code uses for Opus, so the gateway adapts
59
+ those parts of the history for it, as Claude Code itself does. Prompt
60
60
  caching and a claude.ai login work as usual.
61
61
 
62
62
  ## How Jev selects a tier
@@ -76,12 +76,12 @@ flowchart LR
76
76
  class R out
77
77
  ```
78
78
 
79
- | Tier | Model and effort | Jev selects it for |
80
- | -------- | ------------------------- | --------------------------------------------------------------------- |
81
- | `micro` | Haiku 4.5, no thinking | Lookups, trivial edits, one-step mechanical work |
82
- | `low` | Sonnet 5, your effort | Clear, low-risk coding steps with one obvious approach |
83
- | `medium` | Opus 5.5, `high` effort | Features, bug fixes and refactors with interacting constraints |
84
- | `high` | Opus 5.5, `xhigh` effort | Architecture, unclear bugs, security, work where correctness is vital |
79
+ | Tier | Model and effort | Jev selects it for |
80
+ | -------- | -------------------------- | --------------------------------------------------------------------- |
81
+ | `micro` | Haiku 4.5, no thinking | Lookups, trivial edits, one-step mechanical work |
82
+ | `low` | Sonnet 5.5, your effort | Clear, low-risk coding steps with one obvious approach |
83
+ | `medium` | Sonnet 5.5, `xhigh` effort | Features, bug fixes and refactors with interacting constraints |
84
+ | `high` | Opus 5.5, `xhigh` effort | Architecture, unclear bugs, security, work where correctness is vital |
85
85
 
86
86
  The router tells Jev to put correctness before cost, and not to judge by prompt
87
87
  length, language or single topic words. The router then keeps the choice
@@ -61,9 +61,9 @@ flowchart TB
61
61
  - Claude Code builds each request for the model that the alias behaves as
62
62
  (Opus). A model without one of its [request features](configuration.md#models)
63
63
  gets the request without it, the way Claude Code retries after a 400:
64
- - no `mid-conversation-tool-changes` (Sonnet, Haiku): `tool_addition` and
64
+ - no `mid-conversation-tool-changes` (Haiku): `tool_addition` and
65
65
  `tool_removal` blocks go; their cache breakpoint moves to the block before;
66
- - no `per-turn-control` (Sonnet, Haiku): the beta header and the
66
+ - no `per-turn-control` (Haiku): the beta header and the
67
67
  `output_config` on messages (the effort of each turn) go;
68
68
  - no `mid-conversation-system` (Haiku): each `system` message becomes a
69
69
  `<system-reminder>` text in the user message next to it.
@@ -208,6 +208,16 @@ only when that tier is above the current route.
208
208
  - The downgrade bar grows the same way, from the other side: a candidate
209
209
  colder than the incumbent pays a cache write, not just a smaller model.
210
210
  `bar = 0.90 + 0.08 × tax / (tax + $0.50)`. It stays between 0.90 and 0.98.
211
+ - The downgrade tax nets the savings of the next `H` turns
212
+ (`policy.downgradeHorizonTurns`, default 5):
213
+ `tax = max(0, next + (H − 1) × later)`. `next` is the cost difference of
214
+ the next turn, input and output. `later` is the difference of each later
215
+ turn, with both caches warm. Both come from the
216
+ [shadow estimate](#cache-and-cost-model). Example: warm Opus at `xhigh` to
217
+ a cold Sonnet, 104k of context, 4k of output, 1-hour TTL. The input tax is
218
+ $0.40 (bar 0.935); net of five turns it is $0.20 (bar 0.922).
219
+ - When either model has no `output` price, or a zero one, the downgrade tax
220
+ is the input tax alone. A missing price is unknown, not free.
211
221
  - A prompt continues the task when the Jev continuation answer is 0.7 or
212
222
  more. A continuing prompt never votes for a downgrade.
213
223
  - Upgrades have no cooldown. Plan, then execute, then hard work again is a
@@ -260,8 +270,8 @@ tax = max(0, input_cost(candidate) − input_cost(current))
260
270
  `test/fixtures/list-prices.json`, which names its source and date.
261
271
  - **Shadow estimate.** When the Jev choice differs from the current route, the
262
272
  log records the cost of the next turn, the cost of each later turn with
263
- output, and the turns until a switch repays its cost. The policy does not
264
- read it.
273
+ output, and the turns until a switch repays its cost. The downgrade tax
274
+ uses the same arithmetic. The upgrade tax counts input only.
265
275
 
266
276
  ## Modules
267
277
 
@@ -417,7 +427,7 @@ To work on the code:
417
427
 
418
428
  ```sh
419
429
  npm install
420
- git config --local core.hooksPath scripts/git-hooks # Biome, tests and Gitleaks before commit and push
430
+ git config --local core.hooksPath scripts/git-hooks # Biome, ShellCheck, tests and Gitleaks before commit and push
421
431
  npm test && npm run check # node:test, Biome
422
432
  claude --plugin-dir . --model jev-router # with ANTHROPIC_BASE_URL and TYPESAFE_API_KEY
423
433
  ```
@@ -431,7 +441,12 @@ claude --plugin-dir . --model jev-router # with ANTHROPIC_BASE_URL
431
441
  - Without the hint headers, the gateway guesses side requests. A missed side
432
442
  request with a short history counts as a history break.
433
443
  - `x-claude-code-agent-type` goes to the log only.
434
- - The shadow estimate uses the last output size for both routes, but effort
435
- changes how much a model writes.
444
+ - The shadow estimate and the downgrade tax use the last output size for
445
+ both routes, but effort changes how much a model writes.
446
+ - Opus 5.5 and Sonnet 5.5 bind thinking blocks to the model and the
447
+ conversation. Anthropic answers a replay after a prefix change with 400 for
448
+ accounts created on or after 2026-08-31. A switch between the two, with
449
+ thinking blocks in the history, passed on one older account (2026-10-01);
450
+ no newer account was tested.
436
451
 
437
452
  A day of real usage is in [Evaluation](evaluation.md).
@@ -95,12 +95,12 @@ For example, to run `low` on Sonnet at `high` effort and give Jev more time:
95
95
 
96
96
  ### routes
97
97
 
98
- | Tier | Default |
99
- | -------- | ---------------------------------------- |
100
- | `high` | `{ "model": "opus", "effort": "xhigh" }` |
101
- | `medium` | `{ "model": "opus", "effort": "high" }` |
102
- | `low` | `{ "model": "sonnet" }` |
103
- | `micro` | `{ "model": "haiku" }` |
98
+ | Tier | Default |
99
+ | -------- | ------------------------------------------ |
100
+ | `high` | `{ "model": "opus", "effort": "xhigh" }` |
101
+ | `medium` | `{ "model": "sonnet", "effort": "xhigh" }` |
102
+ | `low` | `{ "model": "sonnet" }` |
103
+ | `micro` | `{ "model": "haiku" }` |
104
104
 
105
105
  `model` is a key of `models`. `effort` is `low`, `medium`, `high`, `xhigh` or
106
106
  `max`. Without `effort`, the gateway keeps the effort that Claude Code sent.
@@ -109,16 +109,16 @@ A test makes sure that they agree.
109
109
 
110
110
  ### models
111
111
 
112
- | Alias | `id` | `input` | `output` | `cacheRead` | `contextWindow` | `maxOutput` | `billing` | `efforts` | `features` |
113
- | -------- | ------------------ | ------- | -------- | ----------- | --------------- | ----------- | --------- | --------- | ------------------------- |
114
- | `opus` | `claude-opus-5-5` | 4 | 20 | 0.2 | 1,000,000 | — | `plan` | all five | all three |
115
- | `sonnet` | `claude-sonnet-5` | 2 | 10 | 0.2 | 1,000,000 | — | `plan` | all five | `mid-conversation-system` |
116
- | `haiku` | `claude-haiku-4-5` | 1 | 5 | 0.1 | 200,000 | 64,000 | `plan` | none | none |
112
+ | Alias | `id` | `input` | `output` | `cacheRead` | `contextWindow` | `maxOutput` | `billing` | `efforts` | `features` |
113
+ | -------- | ------------------- | ------- | -------- | ----------- | --------------- | ----------- | --------- | --------- | ---------- |
114
+ | `opus` | `claude-opus-5-5` | 4 | 20 | 0.2 | 1,000,000 | — | `plan` | all five | all three |
115
+ | `sonnet` | `claude-sonnet-5-5` | 2 | 10 | 0.2 | 1,000,000 | — | `plan` | all five | all three |
116
+ | `haiku` | `claude-haiku-4-5` | 1 | 5 | 0.1 | 200,000 | 64,000 | `plan` | none | none |
117
117
 
118
118
  | Field | Meaning |
119
119
  | ------------------------------ | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
120
120
  | `id` | The model id that the gateway sends to Anthropic. |
121
- | `input`, `output`, `cacheRead` | List prices in USD per million tokens. `output` is optional. Only the `shadow` estimate uses it. |
121
+ | `input`, `output`, `cacheRead` | List prices in USD per million tokens. `output` is optional. The `shadow` estimate and the downgrade tax use it. |
122
122
  | `contextWindow` | The window in tokens. A turn goes only to a model that holds the context at 80% fill. |
123
123
  | `maxOutput` | Optional limit for `max_tokens`. Haiku 4.5 rejects more than 64K. |
124
124
  | `billing` | `plan` for the subscription limits. `credits` for models that bill usage credits. |
@@ -133,20 +133,21 @@ from a probe against the real API. A test fails when they differ.
133
133
 
134
134
  The [switching policy](architecture.md#switching-policy) uses these values.
135
135
 
136
- | Key | Default | Meaning |
137
- | --------------------- | ------- | -------------------------------------------------------------------------------------- |
138
- | `upgradeVotes` | `2` | Consecutive votes above the current tier before an upgrade. |
139
- | `upgradeBase` | `0.75` | The upgrade bar when a switch costs nothing. |
140
- | `upgradeSlope` | `0.15` | How much a switching cost can raise the bar: `base + slope × tax / (tax + pivot)`. |
141
- | `upgradePivotUsd` | `0.5` | The switching cost that adds half of the slope. |
142
- | `jumpConfidence` | `0.95` | The confidence for a jump of two tiers without the vote delay. |
143
- | `downgradeVotes` | `2` | Consecutive votes for a lower tier before a downgrade. |
144
- | `downgradeMass` | `0.9` | The downgrade bar when the candidate's cache is already warm. |
145
- | `downgradeSlope` | `0.08` | How much a cold candidate's cache write can raise the bar, same formula as upgrade. |
146
- | `downgradePivotUsd` | `0.5` | The switching cost that adds half of the slope. |
147
- | `continuationMass` | `0.7` | The Jev probability for "continues the task" that keeps the route. |
148
- | `escalationHoldTurns` | `2` | Turns that the route stays up after an escalation. |
149
- | `cashCapUsd` | `2` | The cold-write guard: the largest first cache write for a switch to a `credits` model. |
136
+ | Key | Default | Meaning |
137
+ | ----------------------- | ------- | ----------------------------------------------------------------------------------------------------- |
138
+ | `upgradeVotes` | `2` | Consecutive votes above the current tier before an upgrade. |
139
+ | `upgradeBase` | `0.75` | The upgrade bar when a switch costs nothing. |
140
+ | `upgradeSlope` | `0.15` | How much a switching cost can raise the bar: `base + slope × tax / (tax + pivot)`. |
141
+ | `upgradePivotUsd` | `0.5` | The switching cost that adds half of the slope. |
142
+ | `jumpConfidence` | `0.95` | The confidence for a jump of two tiers without the vote delay. |
143
+ | `downgradeVotes` | `2` | Consecutive votes for a lower tier before a downgrade. |
144
+ | `downgradeMass` | `0.9` | The downgrade bar when the candidate's cache is already warm. |
145
+ | `downgradeSlope` | `0.08` | How much a cold candidate's cache write can raise the bar, same formula as upgrade. |
146
+ | `downgradePivotUsd` | `0.5` | The switching cost that adds half of the slope. |
147
+ | `downgradeHorizonTurns` | `5` | The turns of output and read savings that a downgrade's tax nets. Input only without `output` prices. |
148
+ | `continuationMass` | `0.7` | The Jev probability for "continues the task" that keeps the route. |
149
+ | `escalationHoldTurns` | `2` | Turns that the route stays up after an escalation. |
150
+ | `cashCapUsd` | `2` | The cold-write guard: the largest first cache write for a switch to a `credits` model. |
150
151
 
151
152
  ### cache, jev, context, log
152
153
 
@@ -18,7 +18,7 @@ to read what it does, and how to control it.
18
18
 
19
19
  1. Start Claude Code. `/model` shows `Jev Router (auto)` as the model.
20
20
  2. Send a prompt. If setup added the status line, it shows the route, for
21
- example `jev-router ▸ sonnet-5 · high · same-tier`.
21
+ example `jev-router ▸ sonnet-5-5 · high · same-tier`.
22
22
  3. Run `/router:status`. It shows `Jev routing: active.`
23
23
 
24
24
  If a step fails, read [Troubleshooting](#troubleshooting).
@@ -101,7 +101,7 @@ Router v0.7.0, gateway: http://127.0.0.1:43170, alias `jev-router`.
101
101
  Jev routing: active.
102
102
  Default tier: low.
103
103
  …
104
- Last turn: low → claude-sonnet-5 at high, reason upgrade-pending, context 84210 tokens, cache reads 83904.
104
+ Last turn: low → claude-sonnet-5-5 at high, reason upgrade-pending, context 84210 tokens, cache reads 83904.
105
105
  Why: Jev asked for a higher tier; the route stays until the votes and the mass are enough.
106
106
  Estimate: upgrade mass 0.78 against a bar of 0.84, switching tax $0.66 at list prices, 1 vote(s) in a row, cache: candidate unknown, current warm.
107
107
  ```
@@ -143,7 +143,7 @@ log has no prompt text.
143
143
  | `reason` | The rule that decided. Tool calls show `tool-continuation`. `error`: routing failed and the default tier served. |
144
144
  | `advice` | The Jev probabilities for each tier, and for "continues the task". |
145
145
  | `estimate` | The numbers of a vote: confidence, bar, switching tax, cache state. |
146
- | `shadow` | The cost of the Jev choice against the current route. The policy does not use it. |
146
+ | `shadow` | The cost of the Jev choice against the current route. The downgrade tax uses the same arithmetic. |
147
147
  | `observed` | The usage of the response: model, effort, context tokens, cache reads, output. |
148
148
  | `failed` | Anthropic answered a routed turn with an error: status, tier, model, effort. |
149
149
  | `historyBreak` | A compaction or a rewind. The votes and the cache estimates reset. |
package/lib/config.mjs CHANGED
@@ -14,17 +14,19 @@ export const DEFAULTS = {
14
14
  // the hooks start it again before the next prompt. Two hours outlast /loop wakeups and Monitor waits (up to one
15
15
  // hour), which reach the gateway without a prompt. 0 keeps it running.
16
16
  gateway: { port: 43170, alias: 'jev-router', baselineTier: 'low', auxiliaryTier: 'low', idleShutdownMs: 7_200_000 },
17
- // `high` and `medium` share one model and differ by effort: Opus 5.5 at xhigh is the strongest
18
- // setting this router can ask for, and the tier ladder stays four wide for the policy.
17
+ // Sonnet 5.5 is the worker: `low` at the effort the user picked, `medium` at xhigh, so an escalation from `low`
18
+ // stays on Sonnet's cache instead of a cold write to Opus. `high` is Opus 5.5 at xhigh, the strongest setting this
19
+ // router asks for: Anthropic keeps Opus for open-ended, long-horizon work. xhigh on Sonnet is not yet measured
20
+ // against Opus at high; the `observed` lines in decisions.jsonl are the input for that.
19
21
  routes: {
20
22
  high: { model: 'opus', effort: 'xhigh' },
21
- medium: { model: 'opus', effort: 'high' },
23
+ medium: { model: 'sonnet', effort: 'xhigh' },
22
24
  low: { model: 'sonnet' },
23
25
  micro: { model: 'haiku' },
24
26
  },
25
27
  // `id` is sent upstream verbatim. List prices in USD per million tokens; `cacheRead` is absolute,
26
28
  // not a multiplier (Opus 5.5 reads at 0.05x input, the rest at the standard 0.1x). A price change must also change
27
- // test/fixtures/list-prices.json, with its source and date. `output` feeds only the shadow estimate in the log.
29
+ // test/fixtures/list-prices.json, with its source and date. `output` feeds the shadow estimate and the downgrade tax.
28
30
  // `efforts` lists what the model accepts; an empty list means no effort field and no adaptive thinking. A change
29
31
  // must also change test/fixtures/effort-support.json, from a new probe against the real API.
30
32
  // `features` lists the FEATURES the model accepts; the gateway removes the others from the request. A model
@@ -42,14 +44,14 @@ export const DEFAULTS = {
42
44
  features: FEATURES,
43
45
  },
44
46
  sonnet: {
45
- id: 'claude-sonnet-5',
47
+ id: 'claude-sonnet-5-5',
46
48
  input: 2,
47
49
  output: 10,
48
50
  cacheRead: 0.2,
49
51
  contextWindow: 1_000_000,
50
52
  billing: 'plan',
51
53
  efforts: EFFORTS,
52
- features: ['mid-conversation-system'],
54
+ features: FEATURES,
53
55
  },
54
56
  // `maxOutput` caps `max_tokens`: Claude Code asks for 128K, Haiku 4.5 answers more than 64K with 400.
55
57
  haiku: {
@@ -79,6 +81,7 @@ export const DEFAULTS = {
79
81
  downgradeMass: 0.9,
80
82
  downgradeSlope: 0.08, // a cold candidate raises the bar toward ~0.98, same shape as the upgrade bar
81
83
  downgradePivotUsd: 0.5,
84
+ downgradeHorizonTurns: 5, // turns whose output and read savings offset a downgrade's cache write
82
85
  continuationMass: 0.7,
83
86
  escalationHoldTurns: 2,
84
87
  // Ceiling on a cold cache write to a `credits` model. No default model bills credits, so this is
@@ -213,12 +216,14 @@ function validate(config) {
213
216
  for (const field of ['upgradeBase', 'jumpConfidence', 'downgradeMass', 'continuationMass']) {
214
217
  if (!(isNumber(p[field], 0) && p[field] <= 1)) throw new Error(`policy.${field} must be between 0 and 1`);
215
218
  }
216
- for (const field of ['upgradeVotes', 'downgradeVotes', 'escalationHoldTurns']) {
219
+ for (const field of ['upgradeVotes', 'downgradeVotes', 'downgradeHorizonTurns', 'escalationHoldTurns']) {
217
220
  if (!isNumber(p[field], 1, { integer: true })) throw new Error(`policy.${field} must be a positive integer`);
218
221
  }
219
- if (!isNumber(p.upgradeSlope, 0)) throw new Error('policy.upgradeSlope must be a non-negative number');
220
- // The pivot divides: tax / (tax + pivot). Zero makes a free switch NaN and silently stops upgrades.
221
- if (!isNumber(p.upgradePivotUsd, 0, { above: true })) throw new Error('policy.upgradePivotUsd must be positive');
222
+ // The pivot divides: tax / (tax + pivot). Zero makes a free switch NaN and silently stops that switch.
223
+ for (const side of ['upgrade', 'downgrade']) {
224
+ if (!isNumber(p[`${side}Slope`], 0)) throw new Error(`policy.${side}Slope must be a non-negative number`);
225
+ if (!isNumber(p[`${side}PivotUsd`], 0, { above: true })) throw new Error(`policy.${side}PivotUsd must be positive`);
226
+ }
222
227
  if (!isNumber(p.cashCapUsd, 0)) throw new Error('policy.cashCapUsd must be a non-negative number');
223
228
  const { cache, jev, context } = config;
224
229
  for (const map of ['writeMultiplier', 'ttlMs']) {
package/lib/cost.mjs CHANGED
@@ -2,7 +2,7 @@
2
2
  import { clampEffort } from './rewrite.mjs';
3
3
 
4
4
  // The prompt cache belongs to a model, and its messages part to the effort too: a top-level effort change rewrites
5
- // the messages cache. Opus at `high` and Opus at `xhigh` are two caches, not one.
5
+ // the messages cache. Sonnet at `high` and Sonnet at `xhigh` are two caches, not one.
6
6
  export function cacheKey(modelId, effort) {
7
7
  return effort ? `${modelId}@${effort}` : modelId;
8
8
  }
@@ -62,9 +62,9 @@ export function switchingTaxUsd(config, candidateTier, incumbentTier, facts, now
62
62
  );
63
63
  }
64
64
 
65
- // Shadow estimate for the decision log; the policy does not read it. At list prices, so for `plan` models it is a
66
- // list-price equivalent, not a charge. Output uses the last observed output size for both routes, although effort
67
- // changes how much a model writes.
65
+ // Shadow estimate for the decision log; the downgrade tax nets it over a horizon. At list prices, so for `plan` models
66
+ // it is a list-price equivalent, not a charge. Output uses the last observed output size for both routes, although
67
+ // effort changes how much a model writes.
68
68
  // - nextTurnUsd: candidate minus incumbent for the next request, input at the current cache state plus output.
69
69
  // - laterTurnUsd: the same difference for each later turn, once both caches are warm.
70
70
  // - paybackTurns: 0 when the switch is cheaper at once, n when later turns repay it after n turns, null when never.
@@ -82,3 +82,12 @@ export function shadowEconomics(config, candidateTier, incumbentTier, facts, now
82
82
  else if (nextTurnUsd > 0 && laterTurnUsd < 0) paybackTurns = Math.ceil(nextTurnUsd / -laterTurnUsd);
83
83
  return { nextTurnUsd, laterTurnUsd, paybackTurns, outputTokens };
84
84
  }
85
+
86
+ // A downgrade's tax over `turns` turns: the next one plus the later ones, output included. A missing or zero output
87
+ // price is unknown, not free, so the tax then falls back to the next request's input alone.
88
+ export function downgradeTaxUsd(config, candidateTier, incumbentTier, facts, now, turns) {
89
+ const priced = [candidateTier, incumbentTier].every((tier) => modelOf(config, tier).output > 0);
90
+ if (!priced) return Math.max(0, switchingTaxUsd(config, candidateTier, incumbentTier, facts, now));
91
+ const { nextTurnUsd, laterTurnUsd } = shadowEconomics(config, candidateTier, incumbentTier, facts, now);
92
+ return Math.max(0, nextTurnUsd + (turns - 1) * laterTurnUsd);
93
+ }
package/lib/gateway.mjs CHANGED
@@ -214,7 +214,7 @@ function errorBody(type, message) {
214
214
  return { type: 'error', error: { type, message } };
215
215
  }
216
216
 
217
- // "Picks claude-opus-5-5 / claude-sonnet-5 / claude-haiku-4-5 and the effort for each turn"
217
+ // "Picks claude-opus-5-5 / claude-sonnet-5-5 / claude-haiku-4-5 and the effort for each turn"
218
218
  export function routesDescription(config) {
219
219
  const ids = [...new Set(Object.values(config.routes).map((r) => config.models[r.model].id))];
220
220
  return `Picks ${ids.join(' / ')} and the effort for each turn`;
package/lib/policy.mjs CHANGED
@@ -1,6 +1,14 @@
1
1
  // Switching policy v0 (docs/architecture.md): stickiness, escalation floor, cost-gated votes.
2
2
  import { rank, TIERS } from './config.mjs';
3
- import { cacheState, coldWriteUsd, isWarm, nextContextTokens, routeCacheKey, switchingTaxUsd } from './cost.mjs';
3
+ import {
4
+ cacheState,
5
+ coldWriteUsd,
6
+ downgradeTaxUsd,
7
+ isWarm,
8
+ nextContextTokens,
9
+ routeCacheKey,
10
+ switchingTaxUsd,
11
+ } from './cost.mjs';
4
12
 
5
13
  export function initialState() {
6
14
  return { turn: 0, votes: [], holdUntilTurn: 0, escalatedSignature: null };
@@ -49,8 +57,8 @@ export function decide({ config, facts, advice, state, baseline, now }) {
49
57
  const support = massAtOrBelow(advice.probabilities, choice);
50
58
  const streak = trailing(next.votes, (v) => rank(v.tier) <= rank(choice));
51
59
  // A downgrade is not always cheaper this turn: a cold candidate next to a warm incumbent can pay a bigger cache
52
- // write than it saves. The bar rises with that cost the same way the upgrade bar does.
53
- const tax = Math.max(0, switchingTaxUsd(config, choice, incumbent, facts, now));
60
+ // write than it saves. The bar rises with that cost, net of what the next turns save, the way the upgrade bar does.
61
+ const tax = downgradeTaxUsd(config, choice, incumbent, facts, now, config.policy.downgradeHorizonTurns);
54
62
  const threshold =
55
63
  config.policy.downgradeMass + config.policy.downgradeSlope * (tax / (tax + config.policy.downgradePivotUsd));
56
64
  const estimate = { downgradeMass: support, streak, taxUsd: tax, threshold };
package/lib/router.mjs CHANGED
@@ -185,7 +185,8 @@ export class Router {
185
185
  }
186
186
  const now = this.now();
187
187
  const decision = decide({ config: this.config, facts, advice, state, baseline: gateway.baselineTier, now });
188
- // Shadow only: what following Jev's choice instead of staying would cost at list prices. The policy ignores it.
188
+ // Shadow only: what following Jev's choice instead of staying would cost at list prices. The policy does not read
189
+ // this record; the downgrade tax computes its own from the same function.
189
190
  const incumbent = TIERS.includes(facts.lastRoute) ? facts.lastRoute : gateway.baselineTier;
190
191
  const shadow =
191
192
  TIERS.includes(advice?.choice) && advice.choice !== incumbent
package/package.json CHANGED
@@ -1,6 +1,6 @@
1
1
  {
2
2
  "name": "@alexeiled/claude-router",
3
- "version": "0.7.1",
3
+ "version": "0.8.0",
4
4
  "description": "Claude Code plugin: auto-picks the right model and effort for each turn using Jev routing.",
5
5
  "license": "MIT",
6
6
  "author": "Alexei Ledenev",
@@ -1,6 +1,6 @@
1
1
  #!/usr/bin/env bash
2
- # Staged-only checks: Biome on staged .mjs/.json, Gitleaks on the staged diff.
3
- # Full check/test/pack runs in pre-push.
2
+ # Staged-only checks: Biome on staged .mjs/.json, ShellCheck on staged shell
3
+ # scripts, Gitleaks on the staged diff. Full check/test/pack runs in pre-push.
4
4
  set -euo pipefail
5
5
 
6
6
  cd "$(git rev-parse --show-toplevel)"
@@ -16,12 +16,26 @@ command -v gitleaks >/dev/null || {
16
16
  }
17
17
 
18
18
  files=()
19
+ shell_files=()
19
20
  while IFS= read -r -d '' f; do
20
- case "$f" in *.mjs | *.json) files+=("$f") ;; esac
21
+ case "$f" in
22
+ *.mjs | *.json) files+=("$f") ;;
23
+ *.sh | scripts/git-hooks/*) shell_files+=("$f") ;;
24
+ esac
21
25
  done < <(git diff --cached --name-only -z --diff-filter=ACMR)
22
26
 
23
27
  if [ "${#files[@]}" -gt 0 ]; then
24
28
  "$biome" check --no-errors-on-unmatched -- "${files[@]}"
25
29
  fi
26
30
 
31
+ # Checked only when shell files are staged, so contributors who never touch
32
+ # them don't need shellcheck installed.
33
+ if [ "${#shell_files[@]}" -gt 0 ]; then
34
+ command -v shellcheck >/dev/null || {
35
+ echo "pre-commit: shellcheck missing; run 'brew install shellcheck'" >&2
36
+ exit 1
37
+ }
38
+ shellcheck -- "${shell_files[@]}"
39
+ fi
40
+
27
41
  gitleaks git --pre-commit --staged --redact --verbose --no-banner
@@ -1,8 +1,8 @@
1
1
  ---
2
2
  name: medium
3
3
  disable-model-invocation: true
4
- description: Manual pin: run this turn on opus at high effort. Use as /router:medium <prompt>.
5
- model: opus
6
- effort: high
4
+ description: Manual pin: run this turn on sonnet at xhigh effort. Use as /router:medium <prompt>.
5
+ model: sonnet
6
+ effort: xhigh
7
7
  ---
8
8
  Routing tier applied. Continue with the user's request as written. Do not mention the router.