fix: preserve compaction files and classify Z.AI overflow (#35747)

Co-authored-by: opencode-agent[bot] <219766164+opencode-agent[bot]@users.noreply.github.com>
Co-authored-by: 冯基魁 <56265583+fengjikui@users.noreply.github.com>

packages/core/src/session/compaction.ts

@@ -25,20 +25,27 @@ const SUMMARY_TEMPLATE = `Output exactly the Markdown structure shown inside <te
 - [constraints/preferences, decisions and why, important facts/assumptions, exact context needed to continue, or "(none)"]
 
 ## Work State
-- Completed: [finished work, verified facts, or changes made; otherwise "(none)"]
-- Active: [current work, partial changes, or investigation state; otherwise "(none)"]
-- Blocked: [blockers, failing commands, or unknowns; otherwise "(none)"]
+### Completed
+- [finished work, verified facts, or changes made; otherwise "(none)"]
+
+### Active
+- [current work, partial changes, or investigation state; otherwise "(none)"]
+
+### Blocked
+- [blockers, failing commands, or unknowns; otherwise "(none)"]
 
 ## Next Move
 1. [immediate concrete action, or "(none)"]
 2. [next action if known, or "(none)"]
+
+## Relevant Files
+- [file or directory path: why it matters, or "(none)"]
 </template>
 
 Rules:
 - Keep every section, even when empty.
 - Use terse bullets, not prose paragraphs.
 - Preserve exact file paths, symbols, commands, error strings, URLs, and identifiers when known.
-- Put relevant files and symbols inside the section where they matter; do not add extra sections.
 - Do not mention the summary process or that context was compacted.`
 
 type Settings = {

packages/core/test/session-compaction.test.ts

@@ -52,6 +52,15 @@ const it = testEffect(
   ),
 )
 
+test("compaction prompt preserves detailed work state and relevant files", () => {
+  const prompt = SessionCompaction.buildPrompt({ context: ["conversation history"] })
+
+  expect(prompt).toContain("## Work State\n### Completed")
+  expect(prompt).toContain("### Active")
+  expect(prompt).toContain("### Blocked")
+  expect(prompt).toContain("## Relevant Files")
+})
+
 test("compaction describes tool media without embedding base64", () => {
   const base64 = "iVBORw0KGgoAAAANSUhEUgAAAAEAAAAB"
   const serialized = SessionCompaction.serializeToolContent([
@@ -74,13 +83,14 @@ test("compaction prompt requires the checkpoint headings in order", () => {
     "## Objective",
     "## Important Details",
     "## Work State",
+    "### Completed",
+    "### Active",
+    "### Blocked",
     "## Next Move",
+    "## Relevant Files",
   ])
   expect(prompt).toContain("one or two brief sentences")
   expect(prompt).toContain("constraints/preferences, decisions and why")
-  expect(prompt).toContain("Completed:")
-  expect(prompt).toContain("Active:")
-  expect(prompt).toContain("Blocked:")
   expect(prompt).toContain("immediate concrete action")
   expect(prompt).toContain("next action if known")
   expect(prompt).toContain("Keep every section, even when empty.")

packages/llm/src/provider-error.ts

@@ -6,6 +6,7 @@ const patterns = [
   /input is too long for requested model/i,
   /exceeds the context window/i,
   /input token count.*exceeds the maximum/i,
+  /tokens in request more than max tokens allowed/i,
   /maximum prompt length is \d+/i,
   /reduce the length of the messages/i,
   /maximum context length is \d+ tokens/i,

packages/llm/test/provider-error.test.ts

@@ -0,0 +1,8 @@
+import { describe, expect, test } from "bun:test"
+import { isContextOverflow } from "../src"
+
+describe("provider error classification", () => {
+  test("classifies Z.AI GLM token limit messages as context overflow", () => {
+    expect(isContextOverflow("tokens in request more than max tokens allowed")).toBe(true)
+  })
+})

packages/opencode/test/session/message-v2.test.ts

@@ -1448,6 +1448,7 @@ describe("session.message-v2.fromError", () => {
       "prompt is too long: 213462 tokens > 200000 maximum",
       "Your input exceeds the context window of this model",
       "The input token count (1196265) exceeds the maximum number of tokens allowed (1048575)",
+      "tokens in request more than max tokens allowed",
       "Please reduce the length of the messages or completion",
       "400 status code (no body)",
       "413 status code (no body)",