fix(pi-claude-cli,droid-cli,dashboard): stop ballooning resumed CLI sessions and thread per-chat session

In a multi-iteration tool loop, buildResumePrompt anchored on the last user
message and walked forward through preceding tool results. The only user
message stays at index 0, so each iteration re-sent the entire transcript over
stdin to the resumed CLI session, duplicating the original query plus a
growing stack of tool results into the on-disk session every turn. The agent
saw its own context bloating and called it "compaction." Anchor on the last
assistant message and slice forward instead — the genuine delta since the
last turn already on disk. Same bug+fix in droid-cli (copy of pi-claude-cli).

Quick chat also created a fresh pi/Claude CLI session per user message and
faked continuity by stuffing the last 50 chat_messages into the prompt as
"## Previous Conversation". Replace that with real session continuity:
chat_sessions gains a cliSessionFile column (migration 56); ChatManager
opens the existing pi SessionManager file when present and creates a fresh
one (persisting its path) on the first turn. The prompt now carries only
the new user content.

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>

packages/droid-cli/src/__tests__/prompt-builder.test.ts

@@ -1145,6 +1145,42 @@ describe("buildResumePrompt", () => {
     expect(buildResumePrompt(context)).toBe("Hello from blocks");
   });
 
+  // Regression: multi-iteration tool loops re-anchor on the LAST assistant
+  // turn, not the (only) user message at index 0. Previously this dumped the
+  // entire transcript into a "user" prompt every iteration, ballooning the
+  // resumed session.
+  it("returns ONLY the trailing tool result during a multi-iteration tool loop", () => {
+    const context = {
+      messages: [
+        { role: "user", content: "Find foo" },
+        {
+          role: "assistant",
+          content: [{ type: "toolCall", name: "find", arguments: { pattern: "foo" } }],
+        },
+        { role: "toolResult", toolName: "find", content: "no matches (turn 1)" },
+        {
+          role: "assistant",
+          content: [{ type: "toolCall", name: "find", arguments: { pattern: "foo" } }],
+        },
+        { role: "toolResult", toolName: "find", content: "no matches (turn 2)" },
+      ],
+    };
+    const result = buildResumePrompt(context) as string;
+    expect(result).toContain("no matches (turn 2)");
+    expect(result).not.toContain("no matches (turn 1)");
+    expect(result).not.toContain("Find foo");
+  });
+
+  it("returns empty string mid-loop when only an assistant turn exists since the last delta", () => {
+    const context = {
+      messages: [
+        { role: "user", content: "Hi" },
+        { role: "assistant", content: "Working..." },
+      ],
+    };
+    expect(buildResumePrompt(context)).toBe("");
+  });
+
   it("handles images in the final user message by returning ContentBlock[]", () => {
     const context = {
       messages: [