feat: per-provider/per-model overrides for ALL compress fields via nested compress.providers (issue #344)

CONFIGURATION.md

@@ -157,15 +157,15 @@ Core compression behavior.
 
 #### `compress.maxContextLimit`
 - **Type:** `number | \`${number}%\``
-- **Default:** `"55%"`
+- **Default:** `"80%"`
 - **Status:** ACTIVE
-- **Description:** Upper context usage threshold (as % of model context window or absolute tokens). When exceeded, ACP nudges the model to compress. Example: `"55%"` or `100000`.
+- **Description:** Upper context usage threshold (as % of model context window or absolute tokens). When exceeded, ACP nudges the model to compress. Example: `"80%"` or `100000`.
 
 #### `compress.minContextLimit`
 - **Type:** `number | \`${number}%\``
-- **Default:** `"45%"`
+- **Default:** `"80%"`
 - **Status:** ACTIVE
-- **Description:** Lower context usage threshold. ACP stops nudging when usage drops below this level.
+- **Description:** Lower context usage threshold for turn/iteration reminder nudges. ACP stops injecting those reminders when usage drops below this level (or when the limit cannot be resolved to a concrete value, e.g. a `"X%"` limit with an unknown model context window). Growth nudges are governed separately by `minNudgeContextPercent`.
 
 #### `compress.modelMaxLimits`
 - **Type:** `Record<string, number | \`${number}%\`>`
@@ -187,9 +187,41 @@ Core compression behavior.
 
 #### `compress.minNudgeContextPercent`
 - **Type:** `number`
-- **Default:** `15`
+- **Default:** `5`
+- **Status:** ACTIVE
+- **Description:** Floor for growth-triggered nudges, as a percentage of the model context window: a growth nudge requires context usage at or above this percentage (in addition to the growth threshold). Over-max (`maxContextLimit`) and the 98% emergency-override nudges bypass the floor. If the model context window is unknown, the floor is unresolvable and growth nudges fall back to growth-only behavior. Turn/iteration reminder nudges are governed by `minContextLimit`, not this field. The default is deliberately low: with the default `nudgeGrowthTokens` (50K), a 5% floor stays inert for typical working cycles and only binds on very large (≥2M-class) windows — a higher default (e.g. 15%) would bind on ≥400K windows and shift every compress cycle's working range upward on large-window models. Set `0` to disable the floor entirely, or raise it (e.g. 15–30%) to keep growth nudges waiting until a larger share of the window is in use. Can be narrowed per provider / per model via [`compress.providers`](#compressproviders) (issue #344).
+
+#### `compress.providers`
+- **Type:** `Record<string, ProviderOverrides>` where `ProviderOverrides = Partial<CompressOverridableConfig> & { models?: Record<string, Partial<CompressOverridableConfig>> }` (all fields optional at both levels)
+- **Default:** `undefined`
 - **Status:** ACTIVE
-- **Description:** Minimum context usage percentage before any nudges are shown. Below this, no nudges are injected.
+- **Description:** Nested per-provider / per-model overrides for **every tunable compress field**, resolved field-by-field with the cascade **model > provider > global** (mirrors the sibling project billion-context-pi, issue #344). Deeper levels only override when the field is explicitly set — unset fields never clear shallower values. `0` / `false` are explicit values, not "unset". Unknown provider/model ids fall back to the global value. Percentages and `"X%"` limits resolve against the active model's context window. Across the three config file layers (global → config dir → project) the maps deep-merge per provider/model key — a project layer can narrow one provider without wiping others configured in lower layers.
+- **Overridable fields:** `maxContextLimit`, `emergencyThresholdPercent`, `minNudgeContextPercent`, `nudgeFrequency`, `iterationNudgeThreshold`, `toolOutputNudgeThreshold`, `nudgeGrowthTokens`, `minNudgeGrowthRatio`, `minNudgeGrowthFloor`, `nudgeForce`, `protectedTools`, `showCompression`, `summaryBuffer`, `protectTags`, `protectUserMessages`, `maxSummaryLengthHard`, `minCompressRange`, `maxVisibleSegments`, `keepEmbedMaxChars`, `lastSegmentSoftBlock`, `preserveRecentMessages`, `preserveRecentTokens`, `preserveLastUserMessage`.
+- **Not overridable:** `permission` (session-level, fixed before model info is known), the deprecated `minContextLimit` / `modelMinLimits` family, the flat `modelMaxLimits` map (legacy), and `providers` itself. For `maxContextLimit` the precedence when set nested is **nested override > `modelMaxLimits` flat map > global**. `protectedTools` set here affects the compress tool and nudge-side logic; the system-prompt protected-tools listing (shown at prompt build time, before model info is available) always reflects the global value.
+
+```jsonc
+{
+    "compress": {
+        "maxContextLimit": "55%",
+        "minNudgeContextPercent": 5,
+        "nudgeGrowthTokens": 50000,
+        "providers": {
+            "anthropic": {
+                "minNudgeContextPercent": 8,
+                "nudgeForce": "strong",
+                "models": {
+                    "claude-sonnet-4-6": {
+                        "minNudgeContextPercent": 30,
+                        "maxContextLimit": "70%",
+                        "nudgeGrowthTokens": 20000
+                    }
+                }
+            }
+        }
+    }
+}
+```
+In this example, for `anthropic/claude-sonnet-4-6`: the floor is 30%, the over-max band starts at 70% of the window instead of the global 55% (a larger working range before over-max nudges kick in), the growth threshold is 20K, and nudges use the `strong` tone (inherited from the provider level). Every other Anthropic model gets the 8% floor and `strong` tone but keeps the global band and 50K growth threshold; everything else uses the pure global values. Provider keys are provider ids and model keys are model ids (as reported by the active session, e.g. `anthropic`, `claude-sonnet-4-6`).
 
 #### `compress.nudgeGrowthTokens`
 - **Type:** `number`
@@ -447,6 +479,45 @@ Post-compression quality evaluation. Runs after each compression to verify summa
 }
 ```
 
+### Per-model growth-nudge floor (nested providers.models)
+```jsonc
+{
+    "compress": {
+        "minNudgeContextPercent": 5,
+        "providers": {
+            "anthropic": {
+                "minNudgeContextPercent": 8,
+                "models": {
+                    "claude-sonnet-4-6": { "minNudgeContextPercent": 30 },
+                    "claude-haiku-4-5": { "minNudgeContextPercent": 0 }
+                }
+            }
+        }
+    }
+}
+```
+Floors resolve as model > provider > global (field-by-field). `0` disables the floor for that model — useful for small-window models where the growth threshold alone is the right signal.
+
+### Per-model tuning of any compress field (nested providers.models)
+The same cascade works for every tunable field, not just the floor — e.g. give one heavy model a tighter growth threshold and a lower over-max band while its siblings keep the global profile:
+```jsonc
+{
+    "compress": {
+        "providers": {
+            "anthropic": {
+                "models": {
+                    "claude-sonnet-4-6": {
+                        "nudgeGrowthTokens": 20000,
+                        "maxContextLimit": "40%"
+                    }
+                }
+            }
+        }
+    }
+}
+```
+See the [`compress.providers`](#compressproviders) reference for the full overridable field list.
+
 ### Protect sensitive files
 ```jsonc
 {

CONFIGURATION.zh-CN.md

@@ -187,9 +187,41 @@ ACP 从最多三层配置文件中读取（后加载的覆盖先加载的）：
 
 #### `compress.minNudgeContextPercent`
 - **类型：** `number`
-- **默认值：** `15`
+- **默认值：** `5`
+- **状态：** ACTIVE
+- **说明：** 增长触发型 nudge 的上下文使用率下限（占模型上下文窗口的百分比）：增长 nudge 要求上下文使用率达到或超过此百分比（此外还需满足增长阈值）。超过上限（`maxContextLimit`）与 98% 紧急覆盖的 nudge 不受此下限约束。若模型上下文窗口未知，则无法计算该下限，增长 nudge 回退为仅按增长触发的行为。轮次/迭代提醒 nudge 由 `minContextLimit` 控制，而非此字段。默认值刻意设低：在默认 `nudgeGrowthTokens`（50K）下，5% 下限对典型工作周期不生效，仅在非常大的（≥2M 级）窗口上才生效——更高的默认值（如 15%）会在 ≥400K 窗口上生效，并推高大型窗口模型上每个压缩周期的工作区间。设为 `0` 可完全禁用该下限，或调高（如 15–30%）使增长 nudge 等待更大的窗口占用比例。可通过 [`compress.providers`](#compressproviders) 按 provider / 按模型细化（issue #344）。
+
+#### `compress.providers`
+- **类型：** `Record<string, ProviderOverrides>`，其中 `ProviderOverrides = Partial<CompressOverridableConfig> & { models?: Record<string, Partial<CompressOverridableConfig>> }`（两级均所有字段可选）
+- **默认值：** `undefined`
 - **状态：** ACTIVE
-- **说明：** 触发任何 nudge 的最低上下文使用率百分比。低于此值时不注入 nudge。
+- **说明：** 对**所有可调 compress 字段**的嵌套按 provider / 按模型覆盖，逐字段按 **模型 > provider > 全局** 级联解析（与姊妹项目 billion-context-pi 一致，issue #344）。深层仅在该字段被显式设置时才覆盖——未设置的字段不会清空浅层取值；`0` / `false` 是显式值，而非“未设置”。未知的 provider/model id 回退到全局值。百分比与 `"X%"` 限额按当前激活模型的上下文窗口换算。在三个配置文件层（全局 → 配置目录 → 项目）之间，该映射按 provider/model 键深度合并——项目层可以只细化某个 provider 而不清掉低层配置的其他 provider。
+- **可覆盖字段：** `maxContextLimit`、`emergencyThresholdPercent`、`minNudgeContextPercent`、`nudgeFrequency`、`iterationNudgeThreshold`、`toolOutputNudgeThreshold`、`nudgeGrowthTokens`、`minNudgeGrowthRatio`、`minNudgeGrowthFloor`、`nudgeForce`、`protectedTools`、`showCompression`、`summaryBuffer`、`protectTags`、`protectUserMessages`、`maxSummaryLengthHard`、`minCompressRange`、`maxVisibleSegments`、`keepEmbedMaxChars`、`lastSegmentSoftBlock`、`preserveRecentMessages`、`preserveRecentTokens`、`preserveLastUserMessage`。
+- **不可覆盖：** `permission`（会话级，在得知模型信息前已固定）、已废弃的 `minContextLimit` / `modelMinLimits` 系列、旧版扁平 `modelMaxLimits` 映射、以及 `providers` 本身。`maxContextLimit` 在嵌套层设置时的优先级为 **嵌套覆盖 > `modelMaxLimits` 扁平映射 > 全局**。在此设置的 `protectedTools` 影响压缩工具与 nudge 侧逻辑；系统提示词中的受保护工具列表（在提示词构建时生成，早于模型信息可用）始终反映全局值。
+
+```jsonc
+{
+    "compress": {
+        "maxContextLimit": "55%",
+        "minNudgeContextPercent": 5,
+        "nudgeGrowthTokens": 50000,
+        "providers": {
+            "anthropic": {
+                "minNudgeContextPercent": 8,
+                "nudgeForce": "strong",
+                "models": {
+                    "claude-sonnet-4-6": {
+                        "minNudgeContextPercent": 30,
+                        "maxContextLimit": "70%",
+                        "nudgeGrowthTokens": 20000
+                    }
+                }
+            }
+        }
+    }
+}
+```
+此例中，对 `anthropic/claude-sonnet-4-6`：下限为 30%，超限（over-max）区间从窗口的 70% 开始（而非全局 55%，即更大的工作区间），增长阈值为 20K，nudge 语气为 `strong`（继承自 provider 层）。其他 Anthropic 模型获得 8% 下限与 `strong` 语气，但保留全局的区间与 50K 增长阈值；其余全部使用纯全局值。provider 键为 provider id，model 键为模型 id（取自当前激活会话上报的标识，如 `anthropic`、`claude-sonnet-4-6`）。
 
 #### `compress.nudgeGrowthTokens`
 - **类型：** `number`
@@ -447,6 +479,45 @@ ACP 从最多三层配置文件中读取（后加载的覆盖先加载的）：
 }
 ```
 
+### 按模型设置增长 nudge 下限（嵌套 providers.models）
+```jsonc
+{
+    "compress": {
+        "minNudgeContextPercent": 5,
+        "providers": {
+            "anthropic": {
+                "minNudgeContextPercent": 8,
+                "models": {
+                    "claude-sonnet-4-6": { "minNudgeContextPercent": 30 },
+                    "claude-haiku-4-5": { "minNudgeContextPercent": 0 }
+                }
+            }
+        }
+    }
+}
+```
+下限按 **模型 > provider > 全局** 逐字段解析。`0` 表示为该模型禁用下限——适用于小窗口模型，仅靠增长阈值触发即可。
+
+### 按模型调优任意 compress 字段（嵌套 providers.models）
+同样的级联适用于所有可调字段，而不仅是下限——例如给某个重度使用的模型设置更紧的增长阈值与更低的超限区间，而同 provider 的其他模型保持全局配置：
+```jsonc
+{
+    "compress": {
+        "providers": {
+            "anthropic": {
+                "models": {
+                    "claude-sonnet-4-6": {
+                        "nudgeGrowthTokens": 20000,
+                        "maxContextLimit": "40%"
+                    }
+                }
+            }
+        }
+    }
+}
+```
+完整可覆盖字段列表见 [`compress.providers`](#compressproviders) 参考节。
+
 ### 保护敏感文件
 ```jsonc
 {

README.md

@@ -324,11 +324,12 @@ Each level overrides the previous, so project settings take priority over global
         // Soft upper threshold: above this, ACP keeps injecting strong
         // compression nudges (based on nudgeFrequency), so compression is
         // much more likely. Accepts: number or "X%" of model context window.
-        "maxContextLimit": "55%",
-        // Soft lower threshold for reminder nudges: below this, turn/iteration
-        // reminders are off (compression less likely). At/above this, reminders
-        // are on. Accepts: number or "X%" of model context window.
-        "minContextLimit": "45%",
+        "maxContextLimit": "80%",
+        // Soft lower threshold for turn/iteration reminder nudges: below this,
+        // those reminders are off (compression less likely). At/above this, they
+        // are on. Growth nudges have their own floor: minNudgeContextPercent.
+        // Accepts: number or "X%" of model context window.
+        "minContextLimit": "80%",
         // Optional per-model override for maxContextLimit by providerID/modelID.
         // If present, this wins over the global maxContextLimit.
         // Accepts: number or "X%".
@@ -343,6 +344,21 @@ Each level overrides the previous, so project settings take priority over global
         //     "openai/gpt-5.3-codex": 50000,
         //     "anthropic/claude-sonnet-4.6": "25%"
         // },
+        // Nested per-provider/per-model overrides for ANY compress field
+        // (23 fields: thresholds, nudge behavior, protection, ...).
+        // Resolution is per field: model > provider > global. Unknown
+        // provider/model IDs fall back to the global value.
+        // "providers": {
+        //     "anthropic": {
+        //         "nudgeGrowthTokens": 50000,
+        //         "models": {
+        //             "claude-sonnet-4.6": {
+        //                 "maxContextLimit": "70%",
+        //                 "minNudgeContextPercent": 10
+        //             }
+        //         }
+        //     }
+        // },
         // How often the context-limit nudge fires (1 = every fetch, 5 = every 5th)
         "nudgeFrequency": 5,
         // Start adding compression reminders after this many
@@ -402,6 +418,33 @@ Each level overrides the previous, so project settings take priority over global
 
 </details>
 
+### Per-Provider / Per-Model Overrides
+
+Any `compress` field can be overridden per provider and per model via the nested `compress.providers` map:
+
+```jsonc
+{
+    "compress": {
+        "maxContextLimit": "80%",
+        "providers": {
+            "anthropic": {
+                "nudgeGrowthTokens": 20000,
+                "models": {
+                    "claude-sonnet-4.6": { "maxContextLimit": "70%", "nudgeForce": "strong" }
+                }
+            },
+            "openai": { "nudgeGrowthTokens": 40000 }
+        }
+    }
+}
+```
+
+Resolution is **per field**: model > provider > global. Unknown provider/model IDs fall back to the global value. A nested `maxContextLimit` also wins over the legacy flat `modelMaxLimits` map. Overrides deep-merge across the three config layers (global → config dir → project) per provider/model key.
+
+Not overridable here: `permission`, the deprecated `minContextLimit` family, and the flat `model*Limits` maps themselves.
+
+See the [`compress.providers`](./CONFIGURATION.md#compressproviders) reference in CONFIGURATION.md for the full 23-field list and recipes.
+
 ### Prompt Overrides
 
 ACP exposes six editable prompts:

README.zh-CN.md

@@ -297,6 +297,21 @@ ACP 使用自己的配置文件，按以下顺序搜索：
         //     "openai/gpt-5.3-codex": 50000,
         //     "anthropic/claude-sonnet-4.6": "25%"
         // },
+        // 嵌套的 provider/model 级覆盖：可覆盖任意 compress 字段
+        // （23 项：阈值、nudge 行为、保护策略等）。
+        // 逐字段解析优先级：model > provider > 全局。
+        // 未知的 provider/model ID 回退到全局值。
+        // "providers": {
+        //     "anthropic": {
+        //         "nudgeGrowthTokens": 50000,
+        //         "models": {
+        //             "claude-sonnet-4.6": {
+        //                 "maxContextLimit": "70%",
+        //                 "minNudgeContextPercent": 10
+        //             }
+        //         }
+        //     }
+        // },
         // How often the context-limit nudge fires (1 = every fetch, 5 = every 5th)
         "nudgeFrequency": 5,
         // Start adding compression reminders after this many
@@ -355,6 +370,33 @@ ACP 使用自己的配置文件，按以下顺序搜索：
 
 </details>
 
+### 按 Provider / 按模型覆盖
+
+任意 `compress` 字段都可通过嵌套的 `compress.providers` 按 provider 和按模型覆盖：
+
+```jsonc
+{
+    "compress": {
+        "maxContextLimit": "80%",
+        "providers": {
+            "anthropic": {
+                "nudgeGrowthTokens": 20000,
+                "models": {
+                    "claude-sonnet-4.6": { "maxContextLimit": "70%", "nudgeForce": "strong" }
+                }
+            },
+            "openai": { "nudgeGrowthTokens": 40000 }
+        }
+    }
+}
+```
+
+解析为**逐字段**优先级：model > provider > 全局。未知的 provider/model ID 回退到全局值。嵌套的 `maxContextLimit` 同时优先于旧版扁平 `modelMaxLimits` 映射。多层配置（全局 → 配置目录 → 项目）按 provider/model 键深合并。
+
+不可覆盖：`permission`、已废弃的 `minContextLimit` 系列，以及扁平 `model*Limits` 映射自身。
+
+完整 23 项字段清单与配方见 CONFIGURATION.zh-CN.md 的 [`compress.providers`](./CONFIGURATION.zh-CN.md#compressproviders) 参考节。
+
 ### Prompt 覆盖
 
 ACP 暴露六个可编辑的 prompt：