agentsmesh 0.27.0 → 0.28.0

package/CHANGELOG.md

@@ -1,5 +1,13 @@
 # Changelog
 
+## 0.28.0
+
+### Minor Changes
+
+- e148490: Deterministic pre-edit lesson recall via a `PreToolUse` hook.
+
+  `agentsmesh init --lessons` now wires the recall hook under **both** `PreToolUse` and `PostToolUse`, and `agentsmesh lessons hook` is event-aware (it echoes the harness `hook_event_name`). On targets whose hooks can inject context before a tool call (e.g. Claude Code), matching lessons are now surfaced **before the first edit/command** — guarding the "first touch" the previous PostToolUse-only design left unguarded — with no extra model turn and no compliance dependence. Session dedup keeps each lesson injected at most once per session, and `PostToolUse` remains as the fallback for harnesses that support only post-call context injection. Existing projects pick this up by re-running `agentsmesh init --lessons` (idempotent) and `agentsmesh generate`; behavior is unchanged when no `hook_event_name` is supplied (defaults to `PostToolUse`).
+
 ## 0.27.0
 
 ### Minor Changes

package/README.md

@@ -112,7 +112,7 @@ The memory is one git-tracked file, `.agentsmesh/lessons/lessons.json`, and ever
 agentsmesh init --lessons && agentsmesh generate   # wire the recall/capture loop once
 ```
 
-`init --lessons` drops a small always-on rule into `.agentsmesh/rules/_root.md` (so every target gets the habit), seeds the full operating manual as a `lessons` skill where supported, and wires a `PostToolUse` recall hook on hook-capable tools; agents without shell access use the matching MCP tools (`lessons_query` / `lessons_add`). Because the graph is a normal git-tracked file, a lesson one agent learns today helps every teammate's agent tomorrow, and every change is reviewable like any other diff.
+`init --lessons` drops a small always-on rule into `.agentsmesh/rules/_root.md` (so every target gets the habit), seeds the full operating manual as a `lessons` skill where supported, and wires a recall hook (a `PreToolUse` first-touch guard plus a `PostToolUse` fallback) on hook-capable tools; agents without shell access use the matching MCP tools (`lessons_query` / `lessons_add`). Because the graph is a normal git-tracked file, a lesson one agent learns today helps every teammate's agent tomorrow, and every change is reviewable like any other diff.
 
 Full walkthrough: [Teach your AI agents with lessons](https://samplexbro.github.io/agentsmesh/guides/lessons/) · [`agentsmesh lessons` reference](https://samplexbro.github.io/agentsmesh/cli/lessons/).