Quellcode durchsuchen

Apply PR #21822: refactor: improve compaction to retain recent conversation context

opencode-agent[bot] vor 16 Stunden
Ursprung
Commit
8cf05d65c4

+ 2 - 1
packages/opencode/src/agent/prompt/compaction.txt

@@ -1,6 +1,7 @@
 You are a helpful AI assistant tasked with summarizing conversations.
 
-When asked to summarize, provide a detailed but concise summary of the conversation.
+When asked to summarize, provide a detailed but concise summary of the older conversation history.
+The most recent turns may be preserved verbatim outside your summary, so focus on information that would still be needed to continue the work with that recent context available.
 Focus on information that would be helpful for continuing the conversation, including:
 - What was done
 - What is currently being worked on

+ 14 - 0
packages/opencode/src/config/config.ts

@@ -191,6 +191,20 @@ export const Info = z
       .object({
         auto: z.boolean().optional().describe("Enable automatic compaction when context is full (default: true)"),
         prune: z.boolean().optional().describe("Enable pruning of old tool outputs (default: true)"),
+        tail_turns: z
+          .number()
+          .int()
+          .min(0)
+          .optional()
+          .describe(
+            "Number of recent user turns, including their following assistant/tool responses, to keep verbatim during compaction (default: 2)",
+          ),
+        tail_tokens: z
+          .number()
+          .int()
+          .min(0)
+          .optional()
+          .describe("Token budget for retained recent turn spans during compaction"),
         reserved: z
           .number()
           .int()

+ 123 - 9
packages/opencode/src/session/compaction.ts

@@ -15,7 +15,9 @@ import { NotFoundError } from "@/storage"
 import { ModelID, ProviderID } from "@/provider/schema"
 import { Effect, Layer, Context } from "effect"
 import { InstanceState } from "@/effect"
-import { isOverflow as overflow } from "./overflow"
+import { isOverflow as overflow, usable } from "./overflow"
+import { makeRuntime } from "@/effect/run-service"
+import { fn } from "@/util/fn"
 
 const log = Log.create({ service: "session.compaction" })
 
@@ -31,6 +33,39 @@ export const Event = {
 export const PRUNE_MINIMUM = 20_000
 export const PRUNE_PROTECT = 40_000
 const PRUNE_PROTECTED_TOOLS = ["skill"]
+const DEFAULT_TAIL_TURNS = 2
+const MIN_TAIL_TOKENS = 2_000
+const MAX_TAIL_TOKENS = 8_000
+type Turn = {
+  start: number
+  end: number
+  id: MessageID
+}
+
+function tailBudget(input: { cfg: Config.Info; model: Provider.Model }) {
+  return (
+    input.cfg.compaction?.tail_tokens ??
+    Math.min(MAX_TAIL_TOKENS, Math.max(MIN_TAIL_TOKENS, Math.floor(usable(input) * 0.25)))
+  )
+}
+
+function turns(messages: MessageV2.WithParts[]) {
+  const result: Turn[] = []
+  for (let i = 0; i < messages.length; i++) {
+    const msg = messages[i]
+    if (msg.info.role !== "user") continue
+    if (msg.parts.some((part) => part.type === "compaction")) continue
+    result.push({
+      start: i,
+      end: messages.length,
+      id: msg.info.id,
+    })
+  }
+  for (let i = 0; i < result.length - 1; i++) {
+    result[i].end = result[i + 1].start
+  }
+  return result
+}
 
 export interface Interface {
   readonly isOverflow: (input: {
@@ -84,6 +119,55 @@ export const layer: Layer.Layer<
       return overflow({ cfg: yield* config.get(), tokens: input.tokens, model: input.model })
     })
 
+    const estimate = Effect.fn("SessionCompaction.estimate")(function* (input: {
+      messages: MessageV2.WithParts[]
+      model: Provider.Model
+    }) {
+      const msgs = yield* MessageV2.toModelMessagesEffect(input.messages, input.model)
+      return Token.estimate(JSON.stringify(msgs))
+    })
+
+    const select = Effect.fn("SessionCompaction.select")(function* (input: {
+      messages: MessageV2.WithParts[]
+      cfg: Config.Info
+      model: Provider.Model
+    }) {
+      const limit = input.cfg.compaction?.tail_turns ?? DEFAULT_TAIL_TURNS
+      if (limit <= 0) return { head: input.messages, tail_start_id: undefined }
+      const budget = tailBudget({ cfg: input.cfg, model: input.model })
+      const all = turns(input.messages)
+      if (!all.length) return { head: input.messages, tail_start_id: undefined }
+      const recent = all.slice(-limit)
+      const sizes = yield* Effect.forEach(
+        recent,
+        (turn) =>
+          estimate({
+            messages: input.messages.slice(turn.start, turn.end),
+            model: input.model,
+          }),
+        { concurrency: 1 },
+      )
+      if (sizes.at(-1)! > budget) {
+        log.info("tail fallback", { budget, size: sizes.at(-1) })
+        return { head: input.messages, tail_start_id: undefined }
+      }
+
+      let total = 0
+      let keep: Turn | undefined
+      for (let i = recent.length - 1; i >= 0; i--) {
+        const size = sizes[i]
+        if (total + size > budget) break
+        total += size
+        keep = recent[i]
+      }
+
+      if (!keep || keep.start === 0) return { head: input.messages, tail_start_id: undefined }
+      return {
+        head: input.messages.slice(0, keep.start),
+        tail_start_id: keep.id,
+      }
+    })
+
     // goes backwards through parts until there are PRUNE_PROTECT tokens worth of tool
     // calls, then erases output of older tool calls to free context space
     const prune = Effect.fn("SessionCompaction.prune")(function* (input: { sessionID: SessionID }) {
@@ -146,6 +230,7 @@ export const layer: Layer.Layer<
         throw new Error(`Compaction parent must be a user message: ${input.parentID}`)
       }
       const userMessage = parent.info
+      const compactionPart = parent.parts.find((part): part is MessageV2.CompactionPart => part.type === "compaction")
 
       let messages = input.messages
       let replay:
@@ -176,19 +261,20 @@ export const layer: Layer.Layer<
       const model = agent.model
         ? yield* provider.getModel(agent.model.providerID, agent.model.modelID)
         : yield* provider.getModel(userMessage.model.providerID, userMessage.model.modelID)
+      const cfg = yield* config.get()
+      const history = compactionPart && messages.at(-1)?.info.id === input.parentID ? messages.slice(0, -1) : messages
+      const selected = yield* select({
+        messages: history,
+        cfg,
+        model,
+      })
       // Allow plugins to inject context or replace compaction prompt.
       const compacting = yield* plugin.trigger(
         "experimental.session.compacting",
         { sessionID: input.sessionID },
         { context: [], prompt: undefined },
       )
-      const defaultPrompt = `Provide a detailed prompt for continuing our conversation above.
-Focus on information that would be helpful for continuing the conversation, including what we did, what we're doing, which files we're working on, and what we're going to do next.
-The summary that you construct will be used so that another agent can read it and continue the work.
-Do not call any tools. Respond only with the summary text.
-Respond in the same language as the user's messages in the conversation.
-
-When constructing the summary, try to stick to this template:
+      const defaultPrompt = `When constructing the summary, try to stick to this template:
 ---
 ## Goal
 
@@ -213,7 +299,7 @@ When constructing the summary, try to stick to this template:
 ---`
 
       const prompt = compacting.prompt ?? [defaultPrompt, ...compacting.context].join("\n\n")
-      const msgs = structuredClone(messages)
+      const msgs = structuredClone(selected.head)
       yield* plugin.trigger("experimental.chat.messages.transform", {}, { messages: msgs })
       const modelMessages = yield* MessageV2.toModelMessagesEffect(msgs, model, { stripMedia: true })
       const ctx = yield* InstanceState.context
@@ -276,6 +362,13 @@ When constructing the summary, try to stick to this template:
         return "stop"
       }
 
+      if (compactionPart && selected.tail_start_id && compactionPart.tail_start_id !== selected.tail_start_id) {
+        yield* session.updatePart({
+          ...compactionPart,
+          tail_start_id: selected.tail_start_id,
+        })
+      }
+
       if (result === "continue" && input.auto) {
         if (replay) {
           const original = replay.info
@@ -409,4 +502,25 @@ export const defaultLayer = Layer.suspend(() =>
   ),
 )
 
+const { runPromise } = makeRuntime(Service, defaultLayer)
+
+export async function isOverflow(input: { tokens: MessageV2.Assistant["tokens"]; model: Provider.Model }) {
+  return runPromise((svc) => svc.isOverflow(input))
+}
+
+export async function prune(input: { sessionID: SessionID }) {
+  return runPromise((svc) => svc.prune(input))
+}
+
+export const create = fn(
+  z.object({
+    sessionID: SessionID.zod,
+    agent: z.string(),
+    model: z.object({ providerID: ProviderID.zod, modelID: ModelID.zod }),
+    auto: z.boolean(),
+    overflow: z.boolean().optional(),
+  }),
+  (input) => runPromise((svc) => svc.create(input)),
+)
+
 export * as SessionCompaction from "./compaction"

+ 14 - 0
packages/opencode/src/session/message-v2.ts

@@ -210,6 +210,7 @@ export const CompactionPart = PartBase.extend({
   type: z.literal("compaction"),
   auto: z.boolean(),
   overflow: z.boolean().optional(),
+  tail_start_id: MessageID.zod.optional(),
 }).meta({
   ref: "CompactionPart",
 })
@@ -925,8 +926,21 @@ export function get(input: { sessionID: SessionID; messageID: MessageID }): With
 export function filterCompacted(msgs: Iterable<WithParts>) {
   const result = [] as WithParts[]
   const completed = new Set<string>()
+  let retain: MessageID | undefined
   for (const msg of msgs) {
     result.push(msg)
+    if (retain) {
+      if (msg.info.id === retain) break
+      continue
+    }
+    if (msg.info.role === "user" && completed.has(msg.info.id)) {
+      const part = msg.parts.find((item): item is CompactionPart => item.type === "compaction")
+      if (!part) continue
+      if (!part.tail_start_id) break
+      retain = part.tail_start_id
+      if (msg.info.id === retain) break
+      continue
+    }
     if (msg.info.role === "user" && completed.has(msg.info.id) && msg.parts.some((part) => part.type === "compaction"))
       break
     if (msg.info.role === "assistant" && msg.info.summary && msg.info.finish && !msg.info.error)

+ 13 - 9
packages/opencode/src/session/overflow.ts

@@ -5,18 +5,22 @@ import type { MessageV2 } from "./message-v2"
 
 const COMPACTION_BUFFER = 20_000
 
+export function usable(input: { cfg: Config.Info; model: Provider.Model }) {
+  const context = input.model.limit.context
+  if (context === 0) return 0
+
+  const reserved =
+    input.cfg.compaction?.reserved ?? Math.min(COMPACTION_BUFFER, ProviderTransform.maxOutputTokens(input.model))
+  return input.model.limit.input
+    ? Math.max(0, input.model.limit.input - reserved)
+    : Math.max(0, context - ProviderTransform.maxOutputTokens(input.model))
+}
+
 export function isOverflow(input: { cfg: Config.Info; tokens: MessageV2.Assistant["tokens"]; model: Provider.Model }) {
   if (input.cfg.compaction?.auto === false) return false
-  const context = input.model.limit.context
-  if (context === 0) return false
+  if (input.model.limit.context === 0) return false
 
   const count =
     input.tokens.total || input.tokens.input + input.tokens.output + input.tokens.cache.read + input.tokens.cache.write
-
-  const reserved =
-    input.cfg.compaction?.reserved ?? Math.min(COMPACTION_BUFFER, ProviderTransform.maxOutputTokens(input.model))
-  const usable = input.model.limit.input
-    ? input.model.limit.input - reserved
-    : context - ProviderTransform.maxOutputTokens(input.model)
-  return count >= usable
+  return count >= usable(input)
 }

+ 401 - 4
packages/opencode/test/session/compaction.test.ts

@@ -167,7 +167,19 @@ function layer(result: "continue" | "compact") {
   )
 }
 
-function runtime(result: "continue" | "compact", plugin = Plugin.defaultLayer, provider = ProviderTest.fake()) {
+function cfg(compaction?: Config.Info["compaction"]) {
+  const base = Config.Info.parse({})
+  return Layer.mock(Config.Service)({
+    get: () => Effect.succeed({ ...base, compaction }),
+  })
+}
+
+function runtime(
+  result: "continue" | "compact",
+  plugin = Plugin.defaultLayer,
+  provider = ProviderTest.fake(),
+  config = Config.defaultLayer,
+) {
   const bus = Bus.layer
   return ManagedRuntime.make(
     Layer.mergeAll(SessionCompaction.layer, bus).pipe(
@@ -177,7 +189,7 @@ function runtime(result: "continue" | "compact", plugin = Plugin.defaultLayer, p
       Layer.provide(Agent.defaultLayer),
       Layer.provide(plugin),
       Layer.provide(bus),
-      Layer.provide(Config.defaultLayer),
+      Layer.provide(config),
     ),
   )
 }
@@ -221,7 +233,7 @@ function llm() {
   }
 }
 
-function liveRuntime(layer: Layer.Layer<LLM.Service>, provider = ProviderTest.fake()) {
+function liveRuntime(layer: Layer.Layer<LLM.Service>, provider = ProviderTest.fake(), config = Config.defaultLayer) {
   const bus = Bus.layer
   const status = SessionStatus.layer.pipe(Layer.provide(bus))
   const processor = SessionProcessorModule.SessionProcessor.layer.pipe(Layer.provide(summary))
@@ -236,11 +248,66 @@ function liveRuntime(layer: Layer.Layer<LLM.Service>, provider = ProviderTest.fa
       Layer.provide(Plugin.defaultLayer),
       Layer.provide(status),
       Layer.provide(bus),
-      Layer.provide(Config.defaultLayer),
+      Layer.provide(config),
     ),
   )
 }
 
+function reply(
+  text: string,
+  capture?: (input: LLM.StreamInput) => void,
+): (input: LLM.StreamInput) => Stream.Stream<LLM.Event, unknown> {
+  return (input) => {
+    capture?.(input)
+    return Stream.make(
+      { type: "start" } satisfies LLM.Event,
+      { type: "text-start", id: "txt-0" } satisfies LLM.Event,
+      { type: "text-delta", id: "txt-0", delta: text, text } as LLM.Event,
+      { type: "text-end", id: "txt-0" } satisfies LLM.Event,
+      {
+        type: "finish-step",
+        finishReason: "stop",
+        rawFinishReason: "stop",
+        response: { id: "res", modelId: "test-model", timestamp: new Date() },
+        providerMetadata: undefined,
+        usage: {
+          inputTokens: 1,
+          outputTokens: 1,
+          totalTokens: 2,
+          inputTokenDetails: {
+            noCacheTokens: undefined,
+            cacheReadTokens: undefined,
+            cacheWriteTokens: undefined,
+          },
+          outputTokenDetails: {
+            textTokens: undefined,
+            reasoningTokens: undefined,
+          },
+        },
+      } satisfies LLM.Event,
+      {
+        type: "finish",
+        finishReason: "stop",
+        rawFinishReason: "stop",
+        totalUsage: {
+          inputTokens: 1,
+          outputTokens: 1,
+          totalTokens: 2,
+          inputTokenDetails: {
+            noCacheTokens: undefined,
+            cacheReadTokens: undefined,
+            cacheWriteTokens: undefined,
+          },
+          outputTokenDetails: {
+            textTokens: undefined,
+            reasoningTokens: undefined,
+          },
+        },
+      } satisfies LLM.Event,
+    )
+  }
+}
+
 function wait(ms = 50) {
   return new Promise((resolve) => setTimeout(resolve, ms))
 }
@@ -835,6 +902,210 @@ describe("session.compaction.process", () => {
     })
   })
 
+  test("persists tail_start_id for retained recent turns", async () => {
+    await using tmp = await tmpdir()
+    await Instance.provide({
+      directory: tmp.path,
+      fn: async () => {
+        const session = await svc.create({})
+        await user(session.id, "first")
+        const keep = await user(session.id, "second")
+        await user(session.id, "third")
+        await SessionCompaction.create({
+          sessionID: session.id,
+          agent: "build",
+          model: ref,
+          auto: false,
+        })
+
+        const rt = runtime("continue", Plugin.defaultLayer, wide(), cfg({ tail_turns: 2, tail_tokens: 10_000 }))
+        try {
+          const msgs = await svc.messages({ sessionID: session.id })
+          const parent = msgs.at(-1)?.info.id
+          expect(parent).toBeTruthy()
+          await rt.runPromise(
+            SessionCompaction.Service.use((svc) =>
+              svc.process({
+                parentID: parent!,
+                messages: msgs,
+                sessionID: session.id,
+                auto: false,
+              }),
+            ),
+          )
+
+          const part = (await svc.messages({ sessionID: session.id }))
+            .at(-2)
+            ?.parts.find((item) => item.type === "compaction")
+
+          expect(part?.type).toBe("compaction")
+          if (part?.type === "compaction") expect(part.tail_start_id).toBe(keep.id)
+        } finally {
+          await rt.dispose()
+        }
+      },
+    })
+  })
+
+  test("shrinks retained tail to fit tail token budget", async () => {
+    await using tmp = await tmpdir()
+    await Instance.provide({
+      directory: tmp.path,
+      fn: async () => {
+        const session = await svc.create({})
+        await user(session.id, "first")
+        await user(session.id, "x".repeat(2_000))
+        const keep = await user(session.id, "tiny")
+        await SessionCompaction.create({
+          sessionID: session.id,
+          agent: "build",
+          model: ref,
+          auto: false,
+        })
+
+        const rt = runtime("continue", Plugin.defaultLayer, wide(), cfg({ tail_turns: 2, tail_tokens: 100 }))
+        try {
+          const msgs = await svc.messages({ sessionID: session.id })
+          const parent = msgs.at(-1)?.info.id
+          expect(parent).toBeTruthy()
+          await rt.runPromise(
+            SessionCompaction.Service.use((svc) =>
+              svc.process({
+                parentID: parent!,
+                messages: msgs,
+                sessionID: session.id,
+                auto: false,
+              }),
+            ),
+          )
+
+          const part = (await svc.messages({ sessionID: session.id }))
+            .at(-2)
+            ?.parts.find((item) => item.type === "compaction")
+
+          expect(part?.type).toBe("compaction")
+          if (part?.type === "compaction") expect(part.tail_start_id).toBe(keep.id)
+        } finally {
+          await rt.dispose()
+        }
+      },
+    })
+  })
+
+  test("falls back to full summary when even one recent turn exceeds tail budget", async () => {
+    await using tmp = await tmpdir({ git: true })
+    const stub = llm()
+    let captured = ""
+    stub.push(
+      reply("summary", (input) => {
+        captured = JSON.stringify(input.messages)
+      }),
+    )
+    await Instance.provide({
+      directory: tmp.path,
+      fn: async () => {
+        const session = await svc.create({})
+        await user(session.id, "first")
+        await user(session.id, "y".repeat(2_000))
+        await SessionCompaction.create({
+          sessionID: session.id,
+          agent: "build",
+          model: ref,
+          auto: false,
+        })
+
+        const rt = liveRuntime(stub.layer, wide(), cfg({ tail_turns: 1, tail_tokens: 20 }))
+        try {
+          const msgs = await svc.messages({ sessionID: session.id })
+          const parent = msgs.at(-1)?.info.id
+          expect(parent).toBeTruthy()
+          await rt.runPromise(
+            SessionCompaction.Service.use((svc) =>
+              svc.process({
+                parentID: parent!,
+                messages: msgs,
+                sessionID: session.id,
+                auto: false,
+              }),
+            ),
+          )
+
+          const part = (await svc.messages({ sessionID: session.id }))
+            .at(-2)
+            ?.parts.find((item) => item.type === "compaction")
+
+          expect(part?.type).toBe("compaction")
+          if (part?.type === "compaction") expect(part.tail_start_id).toBeUndefined()
+          expect(captured).toContain("yyyy")
+        } finally {
+          await rt.dispose()
+        }
+      },
+    })
+  })
+
+  test("falls back to full summary when retained tail media exceeds tail budget", async () => {
+    await using tmp = await tmpdir({ git: true })
+    const stub = llm()
+    let captured = ""
+    stub.push(
+      reply("summary", (input) => {
+        captured = JSON.stringify(input.messages)
+      }),
+    )
+    await Instance.provide({
+      directory: tmp.path,
+      fn: async () => {
+        const session = await svc.create({})
+        await user(session.id, "older")
+        const recent = await user(session.id, "recent image turn")
+        await svc.updatePart({
+          id: PartID.ascending(),
+          messageID: recent.id,
+          sessionID: session.id,
+          type: "file",
+          mime: "image/png",
+          filename: "big.png",
+          url: `data:image/png;base64,${"a".repeat(4_000)}`,
+        })
+        await SessionCompaction.create({
+          sessionID: session.id,
+          agent: "build",
+          model: ref,
+          auto: false,
+        })
+
+        const rt = liveRuntime(stub.layer, wide(), cfg({ tail_turns: 1, tail_tokens: 100 }))
+        try {
+          const msgs = await svc.messages({ sessionID: session.id })
+          const parent = msgs.at(-1)?.info.id
+          expect(parent).toBeTruthy()
+          await rt.runPromise(
+            SessionCompaction.Service.use((svc) =>
+              svc.process({
+                parentID: parent!,
+                messages: msgs,
+                sessionID: session.id,
+                auto: false,
+              }),
+            ),
+          )
+
+          const part = (await svc.messages({ sessionID: session.id }))
+            .at(-2)
+            ?.parts.find((item) => item.type === "compaction")
+
+          expect(part?.type).toBe("compaction")
+          if (part?.type === "compaction") expect(part.tail_start_id).toBeUndefined()
+          expect(captured).toContain("recent image turn")
+          expect(captured).toContain("Attached image/png: big.png")
+        } finally {
+          await rt.dispose()
+        }
+      },
+    })
+  })
+
   test("allows plugins to disable synthetic continue prompt", async () => {
     await using tmp = await tmpdir()
     await Instance.provide({
@@ -1195,6 +1466,132 @@ describe("session.compaction.process", () => {
       },
     })
   })
+
+  test("summarizes only the head while keeping recent tail out of summary input", async () => {
+    const stub = llm()
+    let captured = ""
+    stub.push(
+      reply("summary", (input) => {
+        captured = JSON.stringify(input.messages)
+      }),
+    )
+
+    await using tmp = await tmpdir({ git: true })
+    await Instance.provide({
+      directory: tmp.path,
+      fn: async () => {
+        const session = await svc.create({})
+        await user(session.id, "older context")
+        await user(session.id, "keep this turn")
+        await user(session.id, "and this one too")
+        await SessionCompaction.create({
+          sessionID: session.id,
+          agent: "build",
+          model: ref,
+          auto: false,
+        })
+
+        const rt = liveRuntime(stub.layer, wide())
+        try {
+          const msgs = await svc.messages({ sessionID: session.id })
+          const parent = msgs.at(-1)?.info.id
+          expect(parent).toBeTruthy()
+          await rt.runPromise(
+            SessionCompaction.Service.use((svc) =>
+              svc.process({
+                parentID: parent!,
+                messages: msgs,
+                sessionID: session.id,
+                auto: false,
+              }),
+            ),
+          )
+
+          expect(captured).toContain("older context")
+          expect(captured).not.toContain("keep this turn")
+          expect(captured).not.toContain("and this one too")
+          expect(captured).not.toContain("What did we do so far?")
+        } finally {
+          await rt.dispose()
+        }
+      },
+    })
+  })
+
+  test("keeps recent pre-compaction turns across repeated compactions", async () => {
+    const stub = llm()
+    stub.push(reply("summary one"))
+    stub.push(reply("summary two"))
+    await using tmp = await tmpdir()
+    await Instance.provide({
+      directory: tmp.path,
+      fn: async () => {
+        const session = await svc.create({})
+        const u1 = await user(session.id, "one")
+        const u2 = await user(session.id, "two")
+        const u3 = await user(session.id, "three")
+        await SessionCompaction.create({
+          sessionID: session.id,
+          agent: "build",
+          model: ref,
+          auto: false,
+        })
+
+        const rt = liveRuntime(stub.layer, wide(), cfg({ tail_turns: 2, tail_tokens: 10_000 }))
+        try {
+          let msgs = await svc.messages({ sessionID: session.id })
+          let parent = msgs.at(-1)?.info.id
+          expect(parent).toBeTruthy()
+          await rt.runPromise(
+            SessionCompaction.Service.use((svc) =>
+              svc.process({
+                parentID: parent!,
+                messages: msgs,
+                sessionID: session.id,
+                auto: false,
+              }),
+            ),
+          )
+
+          const u4 = await user(session.id, "four")
+          await SessionCompaction.create({
+            sessionID: session.id,
+            agent: "build",
+            model: ref,
+            auto: false,
+          })
+
+          msgs = MessageV2.filterCompacted(MessageV2.stream(session.id))
+          parent = msgs.at(-1)?.info.id
+          expect(parent).toBeTruthy()
+          await rt.runPromise(
+            SessionCompaction.Service.use((svc) =>
+              svc.process({
+                parentID: parent!,
+                messages: msgs,
+                sessionID: session.id,
+                auto: false,
+              }),
+            ),
+          )
+
+          const filtered = MessageV2.filterCompacted(MessageV2.stream(session.id))
+          const ids = filtered.map((msg) => msg.info.id)
+
+          expect(ids).not.toContain(u1.id)
+          expect(ids).not.toContain(u2.id)
+          expect(ids).toContain(u3.id)
+          expect(ids).toContain(u4.id)
+          expect(filtered.some((msg) => msg.info.role === "assistant" && msg.info.summary)).toBe(true)
+          expect(
+            filtered.some((msg) => msg.info.role === "user" && msg.parts.some((part) => part.type === "compaction")),
+          ).toBe(true)
+        } finally {
+          await rt.dispose()
+        }
+      },
+    })
+  })
 })
 
 describe("util.token.estimate", () => {

+ 135 - 1
packages/opencode/test/session/messages-pagination.test.ts

@@ -107,13 +107,14 @@ async function addAssistant(
   return id
 }
 
-async function addCompactionPart(sessionID: SessionID, messageID: MessageID) {
+async function addCompactionPart(sessionID: SessionID, messageID: MessageID, tailStartID?: MessageID) {
   await svc.updatePart({
     id: PartID.ascending(),
     sessionID,
     messageID,
     type: "compaction",
     auto: true,
+    tail_start_id: tailStartID,
   } as any)
 }
 
@@ -780,6 +781,139 @@ describe("MessageV2.filterCompacted", () => {
     })
   })
 
+  test("retains original tail when compaction stores tail_start_id", async () => {
+    await Instance.provide({
+      directory: root,
+      fn: async () => {
+        const session = await svc.create({})
+
+        const u1 = await addUser(session.id, "first")
+        const a1 = await addAssistant(session.id, u1, { finish: "end_turn" })
+        await svc.updatePart({
+          id: PartID.ascending(),
+          sessionID: session.id,
+          messageID: a1,
+          type: "text",
+          text: "first reply",
+        })
+
+        const u2 = await addUser(session.id, "second")
+        const a2 = await addAssistant(session.id, u2, { finish: "end_turn" })
+        await svc.updatePart({
+          id: PartID.ascending(),
+          sessionID: session.id,
+          messageID: a2,
+          type: "text",
+          text: "second reply",
+        })
+
+        const c1 = await addUser(session.id)
+        await addCompactionPart(session.id, c1, u2)
+        const s1 = await addAssistant(session.id, c1, { summary: true, finish: "end_turn" })
+        await svc.updatePart({
+          id: PartID.ascending(),
+          sessionID: session.id,
+          messageID: s1,
+          type: "text",
+          text: "summary",
+        })
+
+        const u3 = await addUser(session.id, "third")
+        const a3 = await addAssistant(session.id, u3, { finish: "end_turn" })
+        await svc.updatePart({
+          id: PartID.ascending(),
+          sessionID: session.id,
+          messageID: a3,
+          type: "text",
+          text: "third reply",
+        })
+
+        const result = MessageV2.filterCompacted(MessageV2.stream(session.id))
+
+        expect(result.map((item) => item.info.id)).toEqual([u2, a2, c1, s1, u3, a3])
+
+        await svc.remove(session.id)
+      },
+    })
+  })
+
+  test("prefers latest compaction boundary when repeated compactions exist", async () => {
+    await Instance.provide({
+      directory: root,
+      fn: async () => {
+        const session = await svc.create({})
+
+        const u1 = await addUser(session.id, "first")
+        const a1 = await addAssistant(session.id, u1, { finish: "end_turn" })
+        await svc.updatePart({
+          id: PartID.ascending(),
+          sessionID: session.id,
+          messageID: a1,
+          type: "text",
+          text: "first reply",
+        })
+
+        const u2 = await addUser(session.id, "second")
+        const a2 = await addAssistant(session.id, u2, { finish: "end_turn" })
+        await svc.updatePart({
+          id: PartID.ascending(),
+          sessionID: session.id,
+          messageID: a2,
+          type: "text",
+          text: "second reply",
+        })
+
+        const c1 = await addUser(session.id)
+        await addCompactionPart(session.id, c1, u2)
+        const s1 = await addAssistant(session.id, c1, { summary: true, finish: "end_turn" })
+        await svc.updatePart({
+          id: PartID.ascending(),
+          sessionID: session.id,
+          messageID: s1,
+          type: "text",
+          text: "summary one",
+        })
+
+        const u3 = await addUser(session.id, "third")
+        const a3 = await addAssistant(session.id, u3, { finish: "end_turn" })
+        await svc.updatePart({
+          id: PartID.ascending(),
+          sessionID: session.id,
+          messageID: a3,
+          type: "text",
+          text: "third reply",
+        })
+
+        const c2 = await addUser(session.id)
+        await addCompactionPart(session.id, c2, u3)
+        const s2 = await addAssistant(session.id, c2, { summary: true, finish: "end_turn" })
+        await svc.updatePart({
+          id: PartID.ascending(),
+          sessionID: session.id,
+          messageID: s2,
+          type: "text",
+          text: "summary two",
+        })
+
+        const u4 = await addUser(session.id, "fourth")
+        const a4 = await addAssistant(session.id, u4, { finish: "end_turn" })
+        await svc.updatePart({
+          id: PartID.ascending(),
+          sessionID: session.id,
+          messageID: a4,
+          type: "text",
+          text: "fourth reply",
+        })
+
+        const result = MessageV2.filterCompacted(MessageV2.stream(session.id))
+
+        expect(result.map((item) => item.info.id)).toEqual([u3, a3, c2, s2, u4, a4])
+
+        await svc.remove(session.id)
+      },
+    })
+  })
+
   test("works with array input", () => {
     // filterCompacted accepts any Iterable, not just generators
     const id = MessageID.ascending()