fix(context-window): Tighten context limits and bound memory excerpts (#67277)

* Tighten context limits and bound memory excerpts

* Align startup context defaults in config docs

* Align qmd memory_get bounds with shared limits

* Preserve qmd partial memory reads

* Fix shared memory read type import

* Add changelog entry for context bounds

docs/.generated/config-baseline.sha256

@@ -1,4 +1,4 @@
-900c26a9b060f1dfa712abfba877bd3bf9c7b0c9f2294faf9834038283ec24b6  config-baseline.json
-d956a1d60f776bba712cb04374a4f5657cad95bb088b536c5e3e4e29d4a21328  config-baseline.core.json
+32d4b07b5a5fbe1c8d299f60b1b9a17c5dc6fc743ec007db212336d7878f125e  config-baseline.json
+48d00213069fa979cacff0e268da241f01c09aa259c19bec86a68dbea4f21bea  config-baseline.core.json
 ef83a06633fc001b5b2535566939186ecb49d05cd1a90b40e54cc58d3e6e44e3  config-baseline.channel.json
 5f5d4e850df6e9854a85b5d008236854ce185c707fdbb566efcf00f8c08b36e3  config-baseline.plugin.json

docs/concepts/system-prompt.md

@@ -177,6 +177,19 @@ and the effective agent skill allowlist when `agents.defaults.skills` or
 
 This keeps the base prompt small while still enabling targeted skill usage.
 
+The skills list budget is owned by the skills subsystem:
+
+- Global default: `skills.limits.maxSkillsPromptChars`
+- Per-agent override: `agents.list[].skillsLimits.maxSkillsPromptChars`
+
+Generic bounded runtime excerpts use a different surface:
+
+- `agents.defaults.contextLimits.*`
+- `agents.list[].contextLimits.*`
+
+That split keeps skills sizing separate from runtime read/injection sizing such
+as `memory_get`, live tool results, and post-compaction AGENTS.md refreshes.
+
 ## Documentation
 
 When available, the system prompt includes a **Documentation** section that points to the

docs/gateway/configuration-reference.md

@@ -988,6 +988,142 @@ Default: `"once"`.
 }
 ```
 
+### Context budget ownership map
+
+OpenClaw has multiple high-volume prompt/context budgets, and they are
+intentionally split by subsystem instead of all flowing through one generic
+knob.
+
+- `agents.defaults.bootstrapMaxChars` /
+  `agents.defaults.bootstrapTotalMaxChars`:
+  normal workspace bootstrap injection.
+- `agents.defaults.startupContext.*`:
+  one-shot `/new` and `/reset` startup prelude, including recent daily
+  `memory/*.md` files.
+- `skills.limits.*`:
+  the compact skills list injected into the system prompt.
+- `agents.defaults.contextLimits.*`:
+  bounded runtime excerpts and injected runtime-owned blocks.
+- `memory.qmd.limits.*`:
+  indexed memory-search snippet and injection sizing.
+
+Use the matching per-agent override only when one agent needs a different
+budget:
+
+- `agents.list[].skillsLimits.maxSkillsPromptChars`
+- `agents.list[].contextLimits.*`
+
+#### `agents.defaults.startupContext`
+
+Controls the first-turn startup prelude injected on bare `/new` and `/reset`
+runs.
+
+```json5
+{
+  agents: {
+    defaults: {
+      startupContext: {
+        enabled: true,
+        applyOn: ["new", "reset"],
+        dailyMemoryDays: 2,
+        maxFileBytes: 16384,
+        maxFileChars: 1200,
+        maxTotalChars: 2800,
+      },
+    },
+  },
+}
+```
+
+#### `agents.defaults.contextLimits`
+
+Shared defaults for bounded runtime context surfaces.
+
+```json5
+{
+  agents: {
+    defaults: {
+      contextLimits: {
+        memoryGetMaxChars: 12000,
+        memoryGetDefaultLines: 120,
+        toolResultMaxChars: 16000,
+        postCompactionMaxChars: 1800,
+      },
+    },
+  },
+}
+```
+
+- `memoryGetMaxChars`: default `memory_get` excerpt cap before truncation
+  metadata and continuation notice are added.
+- `memoryGetDefaultLines`: default `memory_get` line window when `lines` is
+  omitted.
+- `toolResultMaxChars`: live tool-result cap used for persisted results and
+  overflow recovery.
+- `postCompactionMaxChars`: AGENTS.md excerpt cap used during post-compaction
+  refresh injection.
+
+#### `agents.list[].contextLimits`
+
+Per-agent override for the shared `contextLimits` knobs. Omitted fields inherit
+from `agents.defaults.contextLimits`.
+
+```json5
+{
+  agents: {
+    defaults: {
+      contextLimits: {
+        memoryGetMaxChars: 12000,
+        toolResultMaxChars: 16000,
+      },
+    },
+    list: [
+      {
+        id: "tiny-local",
+        contextLimits: {
+          memoryGetMaxChars: 6000,
+          toolResultMaxChars: 8000,
+        },
+      },
+    ],
+  },
+}
+```
+
+#### `skills.limits.maxSkillsPromptChars`
+
+Global cap for the compact skills list injected into the system prompt. This
+does not affect reading `SKILL.md` files on demand.
+
+```json5
+{
+  skills: {
+    limits: {
+      maxSkillsPromptChars: 18000,
+    },
+  },
+}
+```
+
+#### `agents.list[].skillsLimits.maxSkillsPromptChars`
+
+Per-agent override for the skills prompt budget.
+
+```json5
+{
+  agents: {
+    list: [
+      {
+        id: "tiny-local",
+        skillsLimits: {
+          maxSkillsPromptChars: 6000,
+        },
+      },
+    ],
+  },
+}
+```
+
 ### `agents.defaults.imageMaxDimensionPx`
 
 Max pixel size for the longest image side in transcript/tool image blocks before provider calls.

docs/reference/token-use.md

@@ -16,9 +16,12 @@ OpenAI-style models average ~4 characters per token for English text.
 OpenClaw assembles its own system prompt on every run. It includes:
 
 - Tool list + short descriptions
-- Skills list (only metadata; instructions are loaded on demand with `read`)
+- Skills list (only metadata; instructions are loaded on demand with `read`).
+  The compact skills block is bounded by `skills.limits.maxSkillsPromptChars`,
+  with optional per-agent override at
+  `agents.list[].skillsLimits.maxSkillsPromptChars`.
 - Self-update instructions
-- Workspace + bootstrap files (`AGENTS.md`, `SOUL.md`, `TOOLS.md`, `IDENTITY.md`, `USER.md`, `HEARTBEAT.md`, `BOOTSTRAP.md` when new, plus `MEMORY.md` when present or `memory.md` as a lowercase fallback). Large files are truncated by `agents.defaults.bootstrapMaxChars` (default: 20000), and total bootstrap injection is capped by `agents.defaults.bootstrapTotalMaxChars` (default: 150000). `memory/*.md` daily files are not part of the normal bootstrap prompt; they remain on-demand via memory tools on ordinary turns, but bare `/new` and `/reset` can prepend a one-shot startup-context block with recent daily memory for that first turn. That startup prelude is controlled by `agents.defaults.startupContext`.
+- Workspace + bootstrap files (`AGENTS.md`, `SOUL.md`, `TOOLS.md`, `IDENTITY.md`, `USER.md`, `HEARTBEAT.md`, `BOOTSTRAP.md` when new, plus `MEMORY.md` when present or `memory.md` as a lowercase fallback). Large files are truncated by `agents.defaults.bootstrapMaxChars` (default: 12000), and total bootstrap injection is capped by `agents.defaults.bootstrapTotalMaxChars` (default: 60000). `memory/*.md` daily files are not part of the normal bootstrap prompt; they remain on-demand via memory tools on ordinary turns, but bare `/new` and `/reset` can prepend a one-shot startup-context block with recent daily memory for that first turn. That startup prelude is controlled by `agents.defaults.startupContext`.
 - Time (UTC + user timezone)
 - Reply tags + heartbeat behavior
 - Runtime metadata (host/OS/model/thinking)
@@ -36,6 +39,18 @@ Everything the model receives counts toward the context limit:
 - Compaction summaries and pruning artifacts
 - Provider wrappers or safety headers (not visible, but still counted)
 
+Some runtime-heavy surfaces have their own explicit caps:
+
+- `agents.defaults.contextLimits.memoryGetMaxChars`
+- `agents.defaults.contextLimits.memoryGetDefaultLines`
+- `agents.defaults.contextLimits.toolResultMaxChars`
+- `agents.defaults.contextLimits.postCompactionMaxChars`
+
+Per-agent overrides live under `agents.list[].contextLimits`. These knobs are
+for bounded runtime excerpts and injected runtime-owned blocks. They are
+separate from bootstrap limits, startup-context limits, and skills prompt
+limits.
+
 For images, OpenClaw downscales transcript/tool image payloads before provider calls.
 Use `agents.defaults.imageMaxDimensionPx` (default: `1200`) to tune this: