dsh-tdai-memory 0.3.2 → 0.3.4

package/LICENSE

@@ -1,21 +1,21 @@
-MIT License
-
-Copyright (c) 2026 Scorp1o117
-
-Permission is hereby granted, free of charge, to any person obtaining a copy
-of this software and associated documentation files (the "Software"), to deal
-in the Software without restriction, including without limitation the rights
-to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
-copies of the Software, and to permit persons to whom the Software is
-furnished to do so, subject to the following conditions:
-
-The above copyright notice and this permission notice shall be included in all
-copies or substantial portions of the Software.
-
-THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
-IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
-FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
-AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
-LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
-OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
-SOFTWARE.
+MIT License
+
+Copyright (c) 2026 Scorp1o117
+
+Permission is hereby granted, free of charge, to any person obtaining a copy
+of this software and associated documentation files (the "Software"), to deal
+in the Software without restriction, including without limitation the rights
+to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+copies of the Software, and to permit persons to whom the Software is
+furnished to do so, subject to the following conditions:
+
+The above copyright notice and this permission notice shall be included in all
+copies or substantial portions of the Software.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+SOFTWARE.

package/README.md

@@ -11,7 +11,13 @@ Part of the [DeepSeek Harness Enhancement Suite](https://github.com/Scorp1o117/d
 A port of **TencentDB Agent Memory** (Tencent Cloud's open-source four-layer
 memory system, originally an OpenClaw plugin) into DeepSeek Harness.
 
-## Features
+## Compatibility (v0.3.4)
+
+Targets DSH `0.1.7-rc.1` (`next`); npm `latest` is `0.1.5-rc.3`.
+Settings now live in the Profile patch, and the browser uses `configForms`.
+Older hosts require an older plugin release; alpha builds remain `unknown`.
+
+## Features
 
 - **L0 conversation capture**: every turn (turn end, request boundary) is
   written to raw conversation storage (JSONL + SQLite + FTS + vectors)
@@ -50,23 +56,14 @@ Hard-won wiring details:
 
 ## Configuration (profile patch + settings)
 
-Configuration is **settings-namespace driven**: the profile patch is the base
-layer, and the `tdai-memory:` section of `$DSH_HOME/settings.yaml` overrides it
-(LLM/embedding keys live in settings.yaml). The **Web UI Settings → 记忆**
+Configuration is stored in the `tdai-memory` entry of the active Profile patch.
+On first launch, DSH imports the old `$DSH_HOME/settings.yaml` values into that
+entry. The **Web UI Settings → 记忆**
 section edits every field (v0.2.0, write-only keys); TdaiCore is built at
 startup, so changes apply **after a restart**.
 
-```yaml
-# $DSH_HOME/settings.yaml
-tdai-memory:
-  llm:
-    apiKey: 'sk-...'
-  embedding:
-    apiKey: 'local-no-key'
-```
-
-```yaml
-# profile patch (base layer)
+```yaml
+# $DSH_HOME/profiles/web/cordis.patch.yml
 - id: tdai-memory
   name: 'dsh-tdai-memory'
   config:
@@ -97,12 +94,11 @@ then mount it in `$DSH_HOME/profiles/web/cordis.patch.yml`:
 - insert:
     - id: tdai-memory
       name: 'dsh-tdai-memory'
-      config: {}          # keys can live in settings.yaml instead
+      config: {}          # fill through Settings → 记忆
 ```
 
-and restart `dsh web`. LLM/embedding API keys can be set in the Web UI
-settings page (记忆 / Memory) or directly in `settings.yaml` under
-`tdai-memory:`.
+and restart `dsh web`. LLM/embedding API keys can be set in the Web UI
+settings page (记忆 / Memory) or in the `tdai-memory` Profile patch entry.
 
 > **Note for users**
 > - This plugin is a standard **profile bundle** (`dsh.bundle.patch`):

package/README.zh.md

@@ -8,7 +8,13 @@
 
 把 **TencentDB Agent Memory**（腾讯云开源的四层记忆系统，原为 OpenClaw 插件）移植进 DeepSeek Harness。
 
-## 能力
+## 兼容性（v0.3.4）
+
+适配 DSH `0.1.7-rc.1`（npm `next`）；npm `latest` 是 `0.1.5-rc.3`。
+设置保存在当前 Profile patch，Web 端使用 `configForms`。
+旧宿主请使用插件旧版；alpha 构建仍标记 `unknown`。
+
+## 能力
 
 - **L0 对话捕获**：每轮对话（turn 结束、请求边界）自动写入原始对话（JSONL + SQLite + FTS + 向量）
 - **L1 结构化记忆**：后台管线用 LLM 从对话中提取事实/偏好/事件（persona / episodic / instruction），写入 `records/` + SQLite + FTS + 向量

package/THIRD-PARTY-NOTICES.md

@@ -1,20 +1,20 @@
-# Third-party notices
-
-## TencentDB Agent Memory core (vendored in `vendor/tdai`)
-
-`vendor/tdai` is a TypeScript-compiled copy of the host-neutral core of
-[TencentCloud/TencentDB-Agent-Memory](https://github.com/TencentCloud/TencentDB-Agent-Memory)
-(`tdai-memory-openclaw-plugin`, npm: `@tencentdb-agent-memory/memory-tencentdb`),
-compiled with `tsc -p dsh-tsconfig.json` (see the upstream repo for source).
-Its runtime dependencies are declared in this package's `dependencies`.
-
-MIT License — Copyright (c) 2026 TencentCloud
-
-## Runtime dependencies
-
-- `ai` / `@ai-sdk/openai` — Vercel AI SDK (Apache-2.0)
-- `@node-rs/jieba` — Rust jieba bindings (MIT)
-- `sqlite-vec` — vector search extension for SQLite (Apache-2.0 / MIT)
-- `json5` — JSON5 parser (MIT)
-- `undici` — HTTP client (MIT)
-- `@tencentdb-agent-memory/tcvdb-text` — TencentDB text utilities (MIT)
+# Third-party notices
+
+## TencentDB Agent Memory core (vendored in `vendor/tdai`)
+
+`vendor/tdai` is a TypeScript-compiled copy of the host-neutral core of
+[TencentCloud/TencentDB-Agent-Memory](https://github.com/TencentCloud/TencentDB-Agent-Memory)
+(`tdai-memory-openclaw-plugin`, npm: `@tencentdb-agent-memory/memory-tencentdb`),
+compiled with `tsc -p dsh-tsconfig.json` (see the upstream repo for source).
+Its runtime dependencies are declared in this package's `dependencies`.
+
+MIT License — Copyright (c) 2026 TencentCloud
+
+## Runtime dependencies
+
+- `ai` / `@ai-sdk/openai` — Vercel AI SDK (Apache-2.0)
+- `@node-rs/jieba` — Rust jieba bindings (MIT)
+- `sqlite-vec` — vector search extension for SQLite (Apache-2.0 / MIT)
+- `json5` — JSON5 parser (MIT)
+- `undici` — HTTP client (MIT)
+- `@tencentdb-agent-memory/tcvdb-text` — TencentDB text utilities (MIT)

package/client.js

@@ -47,7 +47,7 @@ window.__ModuleLoader__.load({
 
     // ── locale ────────────────────────────────────────────────────────────
     var NS = "tdaiMemory";
-    var inject = ["slots", "locale", "settingsScope", "connection"];
+    var inject = ["slots", "locale", "configForms", "connection"];
     var zh = {
       nav: "记忆",
       intro: "TDAI 记忆配置：L0 捕获 → L1 结构化提取 → 召回注入。密钥只写不读。TdaiCore 在启动时构建，修改后需重启生效。",
@@ -345,7 +345,7 @@ window.__ModuleLoader__.load({
     function apply(ctx) {
       var t = ctx.locale.bind(NS);
       ctx.effect(function () { return ctx.locale.register(NS, { zh: zh, en: en }); }, "dsh-tdai-memory: dictionaries");
-      var scope = ctx.settingsScope.bind({ namespace: "tdai-memory" });
+      var scope = ctx.configForms.get("tdai-memory");
       var api = ctx.connection.api;
       ctx.slots.inject("settings.section", function () {
         return ctx.slots.register({

package/cordis.patch.yml

@@ -1,8 +1,8 @@
-# dsh-tdai-memory — bundle patch layer
-#
-# Declares the plugin as a Cordis loader entry, making the package a standard
-# profile bundle: `dsh plugin --profile <p> add dsh-tdai-memory` installs and
-# mounts it without manual cordis.patch.yml edits.
-- insert:
-    - id: tdai-memory
-      name: 'dsh-tdai-memory'
+# dsh-tdai-memory — bundle patch layer
+#
+# Declares the plugin as a Cordis loader entry, making the package a standard
+# profile bundle: `dsh plugin --profile <p> add dsh-tdai-memory` installs and
+# mounts it without manual cordis.patch.yml edits.
+- insert:
+    - id: tdai-memory
+      name: 'dsh-tdai-memory'

package/index.js

@@ -37,7 +37,7 @@ const Config = z.object({
   /** TDAI data directory; default: ~/.memory-tencentdb/memory-tdai. */
   dataDir: z.string().default(""),
   /** LLM used for L1/L2/L3 extraction (OpenAI-compatible). Falls back to TDAI_LLM_* env. */
-  llm: z.object({
+  llm: z.object({
     baseUrl: z.string().default(""),
     apiKey: z.string().default("").role("secret"),
     model: z.string().default(""),
@@ -52,34 +52,34 @@ const Config = z.object({
     sessionHeaderName: z.string().default("x-opencode-session"),
     /** Fixed session id for LLM requests; empty = persistent auto id. */
     sessionId: z.string().default(""),
-  }),
+  }).default({}),
   /** Embedding endpoint (OpenAI-compatible /v1/embeddings). */
-  embedding: z.object({
+  embedding: z.object({
     baseUrl: z.string().default("http://127.0.0.1:8088/v1"),
     apiKey: z.string().default("").role("secret"),
     model: z.string().default("Qwen3-Embedding-0.6B"),
     dimensions: z.number().default(1024),
     sendDimensions: z.boolean().default(false),
-  }),
+  }).default({}),
   /** Capture conversations into L0. */
   captureEnabled: z.boolean().default(true),
   /** L1 extraction pipeline (facts from conversations). */
-  extraction: z.object({
+  extraction: z.object({
     enabled: z.boolean().default(true),
     /** Conflict detection before storing; extra LLM call, disable if flaky.
      *  Off by default: dedup LLM output parsing is flaky (README matches). */
     enableDedup: z.boolean().default(false),
-  }),
+  }).default({}),
   /** Recall injection settings. */
-  recall: z.object({
+  recall: z.object({
     enabled: z.boolean().default(true),
     maxResults: z.number().default(5),
     scoreThreshold: z.number().default(0.3),
     timeoutMs: z.number().default(3000),
-  }),
+  }).default({}),
   /** Register the two search tools. */
   toolsEnabled: z.boolean().default(true),
-});
+}).volatile();
 
 /**
  * Whether a derived message is real human input.
@@ -390,18 +390,9 @@ function apply(ctx, config) {
   // `installSettingsSection` export (the provider now lives at ctx.settings).
   // Inline the same logic via ctx.inject(["settings"]) — works on both
   // 0.1.1 (module export wrapper) and 0.1.2 (ctx.settings) hosts.
-  ctx.inject(["settings"], (sctx) => {
-    const scope = sctx.settings.register(NS, Config, { base: config });
-    const fire = () => {
-      builtOnce = true;
-      void build(scope.get());
-    };
-    sctx.effect(() => () => {});
-    fire();
-    scope.watch(() => {
-      ctx.logger.warn("[tdai-memory] settings updated; restart to apply (TdaiCore is built at startup)");
-    });
-  });
+  ctx.on("settings/document-updated", (id) => {
+    if (id === NS) ctx.logger.warn("[tdai-memory] settings updated; restart to apply (TdaiCore is built at startup)");
+  });
 
   // Fallback when no settings service ever mounts: build from the entry
   // config. One-shot — if the settings mount arrives later, its build
@@ -410,7 +401,7 @@ function apply(ctx, config) {
   setTimeout(() => {
     if (!builtOnce) {
       builtOnce = true;
-      void build(config);
+      void build(typeof config.get === "function" ? config.get() : config);
     }
   }, 500);
 

package/package.json

@@ -1,78 +1,94 @@
-{
-    "name": "dsh-tdai-memory",
-    "version": "0.3.2",
-    "description": "DeepSeek Harness 的 TencentDB Agent Memory 移植：L0 对话捕获 → L1 结构化记忆提取 → L2 场景/L3 画像，自动召回注入 + 记忆/对话搜索工具；复用现有 ~/.memory-tencentdb/memory-tdai 数据；附 Web UI 设置栏。",
-    "type": "module",
-    "main": "index.js",
-    "exports": {
-        ".": "./index.js",
-        "./client": "./client.js",
-        "./package.json": "./package.json"
-    },
-    "dsh": {
-        "bundle": {
-            "patch": "./cordis.patch.yml"
-        },
-        "client": {
-            "platform": "web"
-        }
-    },
-    "homepage": "https://github.com/Scorp1o117/dsh-tdai-memory",
-    "files": [
-        "index.js",
-        "recall-inject.js",
-        "client.js",
-        "cordis.patch.yml",
-        "vendor",
-        "README.md",
-        "README.zh.md",
-        "LICENSE",
-        "THIRD-PARTY-NOTICES.md"
-    ],
-    "keywords": [
-        "dsh",
-        "dsh-plugin",
-        "deepseek-harness",
-        "memory",
-        "tencentdb",
-        "agent-memory",
-        "rag",
-        "vector-search",
-        "persona"
-    ],
-    "license": "MIT",
-    "engines": {
-        "node": ">=22.18"
-    },
-    "repository": {
-        "type": "git",
-        "url": "git+https://github.com/Scorp1o117/dsh-tdai-memory.git"
-    },
-    "bugs": {
-        "url": "https://github.com/Scorp1o117/dsh-tdai-memory/issues"
-    },
-    "dependencies": {
-        "@ai-sdk/openai": "^3.0.53",
-        "@deepseek-ai/schemastery": "^3.18.1",
-        "@node-rs/jieba": "^2.0.1",
-        "@tencentdb-agent-memory/tcvdb-text": "^0.1.1",
-        "ai": "^6.0.164",
-        "json5": "^2.2.3",
-        "sqlite-vec": "0.1.7-alpha.2",
-        "undici": "^8.1.0"
-    },
-    "peerDependencies": {
-        "react": "^18.2.0",
-        "@deepseek-ai/dsh-tools": "^0.1.0-rc.7 || ^0.1.1-rc.1 || ^0.1.2-rc.1 || ^0.1.5-rc.1",
-        "@deepseek-ai/dsh-settings": "^0.1.0-rc.7 || ^0.1.1-rc.1 || ^0.1.2-rc.1 || ^0.1.5-rc.1",
-        "node-llama-cpp": "^3.19.1"
-    },
-    "peerDependenciesMeta": {
-        "node-llama-cpp": {
-            "optional": true
-        }
-    },
-    "scripts": {
-        "test": "node tests/settings-scope-compat.test.mjs && node tests/install-contract.test.mjs"
-    }
-}
+{
+    "name": "dsh-tdai-memory",
+    "version": "0.3.4",
+    "description": "DeepSeek Harness 的 TencentDB Agent Memory 移植：L0 对话捕获 → L1 结构化记忆提取 → L2 场景/L3 画像，自动召回注入 + 记忆/对话搜索工具；复用现有 ~/.memory-tencentdb/memory-tdai 数据；附 Web UI 设置栏。",
+    "type": "module",
+    "main": "index.js",
+    "exports": {
+        ".": "./index.js",
+        "./client": "./client.js",
+        "./package.json": "./package.json"
+    },
+    "dsh": {
+        "bundle": {
+            "patch": "./cordis.patch.yml"
+        },
+        "client": {
+            "platform": "web"
+        },
+        "compatibility": {
+            "dsh": ">=0.1.7-rc.1 <0.2.0",
+            "dshReleases": {
+                "0.1.5-rc.2": "incompatible",
+                "0.1.5-rc.3": "incompatible",
+                "0.1.6-alpha.1": "unknown",
+                "0.1.6-alpha.2": "unknown",
+                "0.1.7-alpha.1": "unknown",
+                "0.1.7-alpha.2": "unknown",
+                "0.1.7-rc.1": "compatible"
+            },
+            "node": ">=22.18",
+            "profiles": [
+                "web"
+            ]
+        }
+    },
+    "homepage": "https://github.com/Scorp1o117/dsh-tdai-memory",
+    "files": [
+        "index.js",
+        "recall-inject.js",
+        "client.js",
+        "cordis.patch.yml",
+        "vendor",
+        "README.md",
+        "README.zh.md",
+        "LICENSE",
+        "THIRD-PARTY-NOTICES.md"
+    ],
+    "keywords": [
+        "dsh",
+        "dsh-plugin",
+        "deepseek-harness",
+        "memory",
+        "tencentdb",
+        "agent-memory",
+        "rag",
+        "vector-search",
+        "persona"
+    ],
+    "license": "MIT",
+    "engines": {
+        "node": ">=22.18"
+    },
+    "repository": {
+        "type": "git",
+        "url": "git+https://github.com/Scorp1o117/dsh-tdai-memory.git"
+    },
+    "bugs": {
+        "url": "https://github.com/Scorp1o117/dsh-tdai-memory/issues"
+    },
+    "dependencies": {
+        "@ai-sdk/openai": "^3.0.53",
+        "@deepseek-ai/schemastery": "^3.18.4",
+        "@node-rs/jieba": "^2.0.1",
+        "@tencentdb-agent-memory/tcvdb-text": "^0.1.1",
+        "ai": "^6.0.164",
+        "json5": "^2.2.3",
+        "sqlite-vec": "0.1.7-alpha.2",
+        "undici": "^8.1.0"
+    },
+    "peerDependencies": {
+        "react": "^18.2.0",
+        "@deepseek-ai/dsh-tools": "^0.1.7-rc.1",
+        "@deepseek-ai/dsh-settings": "^0.1.7-rc.1",
+        "node-llama-cpp": "^3.19.1"
+    },
+    "peerDependenciesMeta": {
+        "node-llama-cpp": {
+            "optional": true
+        }
+    },
+    "scripts": {
+        "test": "node tests/settings-scope-compat.test.mjs && node tests/install-contract.test.mjs"
+    }
+}

package/recall-inject.js

@@ -1,86 +1,86 @@
-/**
- * tdai-recall-inject — agent-plane recall injection row for dsh-tdai-memory.
- *
- * Mounted INSIDE an agent preset (agent.cordis.yml), this row listens on the
- * `system-prompt/assemble` waterfall in the preset's scope, so it runs for
- * every assembly of every agent joined to that preset (agent -> preset scope
- * chain). It asks the `tdaiMemory` service (provided by the root-level
- * dsh-tdai-memory plugin) for relevant long-term memories and appends them
- * as dynamic context sections (user-role snapshots):
- *
- *   - tdai:recall  — relevant L1 memories for the current user message
- *   - tdai:profile — stable persona / scene navigation context
- *
- * The recall call is bounded by `recall.timeoutMs`; failures degrade to no
- * injection and never break the assembly.
- */
-import z from "@deepseek-ai/schemastery";
-
-/** Cordis plugin name. */
-const name = "tdai-recall-inject";
-/** Services consumed: the prompt registry and the memory core service. */
-const inject = ["systemPrompt", "tdaiMemory"];
-
-/** Runtime schema for the recall-inject row. */
-const Config = z.object({
-  /** Skip recall when the user message is shorter than this. */
-  minUserTextChars: z.number().default(1),
-  /** Per-session cache TTL for identical user text (ms). */
-  cacheTtlMs: z.number().default(30_000),
-});
-
-function apply(ctx, config) {
-  const cache = new Map();
-
-  ctx.on("system-prompt/assemble", async (assembly, context) => {
-    try {
-      const agent = context?.agent;
-      if (!agent) return assembly;
-      const session = agent.session;
-      if (!session) return assembly;
-      // Only real human input triggers recall. DSH injects synthetic
-      // user-role context (system reminders, runtime context, skill
-      // catalogs, …) whose source.kind is not "user"; recalling on those
-      // would pollute the prompt with irrelevant memories (issue #1).
-      const lastUser = [...session.deriveMessages()]
-        .reverse()
-        .find((m) => m.role === "user" && (m.source == null || m.source.kind === "user"));
-      if (!lastUser) return assembly;
-      const text = extractText(lastUser.content);
-      if (!text || text.length < config.minUserTextChars) return assembly;
-
-      const now = Date.now();
-      const cached = cache.get(session.id);
-      if (cached && cached.text === text && now - cached.ts < config.cacheTtlMs) {
-        return injectRecall(assembly, cached.result);
-      }
-      const result = await ctx.tdaiMemory
-        .handleBeforeRecall(text, session.id)
-        .catch(() => null);
-      cache.set(session.id, { text, result: result ?? {}, ts: now });
-      return injectRecall(assembly, result ?? {});
-    } catch {
-      return assembly;
-    }
-  });
-}
-
-function extractText(content) {
-  if (typeof content === "string") return content.trim();
-  if (!Array.isArray(content)) return "";
-  return content
-    .filter((b) => b && b.type === "text" && typeof b.text === "string")
-    .map((b) => b.text)
-    .join("\n")
-    .trim();
-}
-
-function injectRecall(assembly, result) {
-  const extras = [];
-  if (result.prependContext) extras.push({ name: "tdai:recall", order: 1000, text: result.prependContext });
-  if (result.appendSystemContext) extras.push({ name: "tdai:profile", order: 1001, text: result.appendSystemContext });
-  if (extras.length === 0) return assembly;
-  return { ...assembly, contexts: [...assembly.contexts, ...extras] };
-}
-
-export { Config, apply, inject, name };
+/**
+ * tdai-recall-inject — agent-plane recall injection row for dsh-tdai-memory.
+ *
+ * Mounted INSIDE an agent preset (agent.cordis.yml), this row listens on the
+ * `system-prompt/assemble` waterfall in the preset's scope, so it runs for
+ * every assembly of every agent joined to that preset (agent -> preset scope
+ * chain). It asks the `tdaiMemory` service (provided by the root-level
+ * dsh-tdai-memory plugin) for relevant long-term memories and appends them
+ * as dynamic context sections (user-role snapshots):
+ *
+ *   - tdai:recall  — relevant L1 memories for the current user message
+ *   - tdai:profile — stable persona / scene navigation context
+ *
+ * The recall call is bounded by `recall.timeoutMs`; failures degrade to no
+ * injection and never break the assembly.
+ */
+import z from "@deepseek-ai/schemastery";
+
+/** Cordis plugin name. */
+const name = "tdai-recall-inject";
+/** Services consumed: the prompt registry and the memory core service. */
+const inject = ["systemPrompt", "tdaiMemory"];
+
+/** Runtime schema for the recall-inject row. */
+const Config = z.object({
+  /** Skip recall when the user message is shorter than this. */
+  minUserTextChars: z.number().default(1),
+  /** Per-session cache TTL for identical user text (ms). */
+  cacheTtlMs: z.number().default(30_000),
+});
+
+function apply(ctx, config) {
+  const cache = new Map();
+
+  ctx.on("system-prompt/assemble", async (assembly, context) => {
+    try {
+      const agent = context?.agent;
+      if (!agent) return assembly;
+      const session = agent.session;
+      if (!session) return assembly;
+      // Only real human input triggers recall. DSH injects synthetic
+      // user-role context (system reminders, runtime context, skill
+      // catalogs, …) whose source.kind is not "user"; recalling on those
+      // would pollute the prompt with irrelevant memories (issue #1).
+      const lastUser = [...session.deriveMessages()]
+        .reverse()
+        .find((m) => m.role === "user" && (m.source == null || m.source.kind === "user"));
+      if (!lastUser) return assembly;
+      const text = extractText(lastUser.content);
+      if (!text || text.length < config.minUserTextChars) return assembly;
+
+      const now = Date.now();
+      const cached = cache.get(session.id);
+      if (cached && cached.text === text && now - cached.ts < config.cacheTtlMs) {
+        return injectRecall(assembly, cached.result);
+      }
+      const result = await ctx.tdaiMemory
+        .handleBeforeRecall(text, session.id)
+        .catch(() => null);
+      cache.set(session.id, { text, result: result ?? {}, ts: now });
+      return injectRecall(assembly, result ?? {});
+    } catch {
+      return assembly;
+    }
+  });
+}
+
+function extractText(content) {
+  if (typeof content === "string") return content.trim();
+  if (!Array.isArray(content)) return "";
+  return content
+    .filter((b) => b && b.type === "text" && typeof b.text === "string")
+    .map((b) => b.text)
+    .join("\n")
+    .trim();
+}
+
+function injectRecall(assembly, result) {
+  const extras = [];
+  if (result.prependContext) extras.push({ name: "tdai:recall", order: 1000, text: result.prependContext });
+  if (result.appendSystemContext) extras.push({ name: "tdai:profile", order: 1001, text: result.appendSystemContext });
+  if (extras.length === 0) return assembly;
+  return { ...assembly, contexts: [...assembly.contexts, ...extras] };
+}
+
+export { Config, apply, inject, name };