@useago/sdk 1.5.2 → 1.5.4

package/CHANGELOG.md

@@ -5,6 +5,24 @@ All notable changes to `@useago/sdk` are documented here.
 The format is based on [Keep a Changelog](https://keepachangelog.com/en/1.1.0/),
 and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0.html).
 
+## [1.5.4] - 2026-07-22
+
+### Added
+
+- `AgoClient.recordActivity(entry)`: record a user (or agent) action into an
+  activity ledger so the agent knows what just happened. The recent window (last
+  10 by default) rides along as the `activity:recent` context entry; the agent's
+  own client-function calls are recorded automatically. Read or clear with
+  `getRecentActivity()` / `clearActivity()`, observe via the `activity:recorded`
+  event, or use the React `useAgoActivityLog()` hook.
+- `AgoConfig.maxActivityEntries`: override how many recent actions the activity
+  ledger keeps (default `10`). Each entry's `data` is size-clamped before storage
+  (long strings truncated, large arrays capped) so a single event can't bloat the
+  context sent on every message.
+- Page-state changes are now reported to the agent: each message includes a
+  `state:delta` entry listing the page-state fields that changed since the last
+  message. A navigation resets the baseline instead of reporting a diff.
+
 ## [1.5.2] - 2026-07-13
 
 ### Added

package/README.md

@@ -18,6 +18,16 @@ orchestration, evaluations, quality monitoring, security controls, human
 fallback, and continuous improvement. Your team ships product-native agents that
 resolve customer needs, without rebuilding the full customer-facing agent platform.
 
+> [!TIP]
+> **Using Claude Code, Codex, or Cursor?** You don't have to read these docs.
+> The full documentation lives in a single file built for coding agents. Paste
+> this into your agent and let it do the integration:
+>
+> ```text
+> Read https://raw.githubusercontent.com/useago/ago-sdk/refs/heads/main/llms-full.txt
+> and integrate the AGO chat SDK into this app.
+> ```
+
 In two minutes you'll have a working chat on your page and an agent that can
 navigate your app. The quickstart below uses **React**; every other stack gets
 the same two steps through its own guide: