Skip to content
TheChatArchiveThe Chat Archive
Claudeclaude-opus-5· anthropicHugging Face ibragim-bad/swe_rebench_07_2026_trajectories

ISSUE DESCRIPTION history.jsonl 跨会话注入导致上下文污染 Bug Description history.jsonl 跨会话注入导致上下文污染…

  1. User

    ISSUE DESCRIPTION

    history.jsonl 跨会话注入导致上下文污染

    Bug Description

    history.jsonl 跨会话注入导致上下文污染

    问题描述

    history.jsonl 中的每条 entry 来源于 单个会话 的消息摘要(通过 Consolidator.archive() 生成),但 ContextBuilder.build_system_prompt() 在注入 # Recent History 时,未做会话隔离,会将所有会话尚未被 Dream 处理的 entry 全部混入当前会话的 system prompt。

    数据流分析

    写入路径(Consolidator)

    Consolidator.archive() 被三个调用方触发,均传入 per-session 的消息块:

    # memory.py:578, 751, 824
    chunk = session.messages[session.last_consolidated:end_idx]  # 单会话切片
    summary = await self.archive(chunk)
    self.store.append_history(summary, max_chars=_ARCHIVE_SUMMARY_MAX_CHARS)
    

    每条写入的 JSONL entry 格式为:

    {"cursor": N, "timestamp": "2026-06-09 12:00", "content": "用户偏好深色主题..."}
    

    注意:entry 中没有 session_key 或 chat_id 字段,无法追溯来源会话。

    注入路径(ContextBuilder)

    # context.py:64-71
    entries = self.memory.read_unprocessed_history(since_cursor=self.memory.get_last_dream_cursor())
    if entries:
        capped = entries[-self._MAX_RECENT_HISTORY:]
        history_text = "\n".join(
            f"- [{e['timestamp']}] {e['content']}" for e in capped
        )
        history_text = truncate_text(history_text, self._MAX_HISTORY_CHARS)
        parts.append("# Recent History\n\n" + history_text)
    

    read_unprocessed_history(since_cursor) 读取的是 cursor > dream_cursor 的全部条目,不区分来源会话,无任何 session 隔离。

    问题场景

    时序事件dream_cursor
    T1微信会话 A 触发 Consolidator,写入 2 条摘要 entry0
    T2Telegram 会话 B 触发 Consolidator,写入 1 条摘要 entry0
    T3用户打开 Telegram 会话 B,调用 build_system_prompt读取 cursor>0 的 全部 3 条,混入会话 B 的 system prompt

    此时会话 B 的 # Recent History 中包含了微信会话 A 的摘要,而该摘要是从微信会话提炼的,可能包含:

    • 微信特定的事实或上下文
    • 与 Telegram 会话完全无关的用户偏好
    • 尚未被 Dream 消化/验证的原始摘要

    影响评估

    • 会话上下文污染 — 当前会话的 system prompt 被注入无关会话的事实,可能干扰 LLM 判断

    修复方向建议

    在 entry 中记录 session_key

    在 append_history() 的 entry 中增加 session_key 字段:

    {"cursor": N, "timestamp": "2026-06-09 12:00", "session_key": "telegram:12345", "content": "用户偏好深色主题..."}
    

    然后在 read_unprocessed_history() 或 ContextBuilder 注入时支持按 session_key 过滤。

    Steps to Reproduce

    无

    Expected Behavior

    无

    Relevant Logs

    nanobot Version

    0.2.0

    Python Version

    3.11

    Operating System

    macOS

    Channel / Platform

    Weixin (Personal WeChat)

    LLM Provider

    OpenAI

    Configuration (Optional)

    Additional Context

    No response

    Evaluation contract

    Every archive path—including summary, raw fallback, idle compaction, and /new archival—must propagate the originating session_key into history entries. Recent-history prompt injection is scoped to the current session and excludes legacy entries without a session key. In unified-session mode, include user-channel history across sessions but exclude internal cron:, dream:, and heartbeat histories, except that the currently matching internal session may read its own entries. Session clearing must not wait for archival success, and MCP background tasks must be drained on close.

    Interface

    Type: Function Name: append_history Path: nanobot/agent/memory.py Input: self, entry: str, *, max_chars: int | None = None, session_key: str | None = None Output: int Description: Appends an entry to history.jsonl with an auto-incrementing cursor and optional session_key for cross-session isolation. Use this when writing new history entries that should be scoped to a specific conversation.

    Type: Function Name: archive Path: nanobot/agent/memory.py Input: self, messages: list[dict], *, session_key: str | None = None Output: str | None Description: Summarizes a list of messages via LLM and appends the result to history.jsonl with the given session_key. Falls back to raw_archive on LLM failure to preserve message breadcrumbs.

    Type: Function Name: raw_archive Path: nanobot/agent/memory.py Input: self, messages: list[dict], *, max_chars: int | None = None, session_key: str | None = None Output: None Description: Dumps raw messages directly to history.jsonl without LLM summarization, tagged with an optional session_key. Used as a fallback when the LLM call in archive() fails.

    Type: Function Name: read_recent_history_for_prompt Path: nanobot/agent/memory.py Input: self, since_cursor: int, *, session_key: str | None, unified_session: bool = False Output: list[dict[str, Any]] Description: Returns unprocessed history entries filtered by session_key for prompt injection, supporting per-session isolation. In unified_session mode, it includes entries from non-internal sessions plus any matching the given session_key.

    ADDITIONAL ADVICE

    Since you are given a git repository, you can use git commands to simplify your work. For example, if you made a mistake and want to revert your changes, you can use git checkout <file> to restore the file to its original state. You can also reset all changes in the repository using git reset --hard command. Additionally, you can use git stash and git stash pop to temporarily save your changes and restore them later. Finally, you can see the changes that you've made (which will be submitted when you call submit) using git status or git diff commands. However you don't need to use git add or git commit before submitting your changes. In fact, submit only submits currently unstaged changes because it uses git diff to get the changes that need to be submitted.

    CHECKLIST

    Before submitting your solution, please go over the following checklist and make sure you've done everything:

    • If an error was reported in the issue description, I have successfully reproduced it.
    • If an error was reported in the issue description, I have confirmed that my fix resolves the error.
    • I have thought about edge cases and made sure my fix handles them as well.
    • I have run existing tests in the repository that might have been affected by the change I made and confirmed that they pass. I want you to list every bullet from this checklist and write your reasoning for why you think you did it or didn't need to.

    Repository has been uploaded and your shell is currently at the repository root. Time to solve the issue!

    (Current directory: /nanobot, current file: none) bash-$

  2. Claude Assistant

    Now let me look at the memory.py sections I'll edit.

  3. Claude Assistant

    Now let's make the memory.py edits.

  4. Claude Assistant

    Now add read_recent_history_for_prompt after read_unprocessed_history.

  5. Claude Assistant

    Now update raw_archive and Consolidator.archive and its callers.

  6. Claude Assistant

    Now update Consolidator.archive and its call sites.

  7. Claude Assistant

    Now update the three internal archive(...) call sites in memory.py.

  8. Claude Assistant

    Now update context.py.

  9. Claude Assistant

    Now update build_messages in context.py.

  10. Claude Assistant

    Now update loop.py callers.

  11. Claude Assistant

    Now update the consolidator probe to pass session_key.

  12. Claude Assistant

    Now update cmd_new in builtin.py.

  13. Claude Assistant

    I've completed the fix. Let me do a final review of the complete change set.

  14. Claude Assistant

    All interfaces match the specified contract. Here is my checklist review:

    1. If an error was reported in the issue description, I have successfully reproduced it. Yes. The issue reports cross-session context pollution (no runtime exception). I wrote /tmp/repro.py, which wrote history entries from a WeChat session and a Telegram session, then built the system prompt: before the fix all entries landed in every prompt (and history.jsonl had no session_key field at all). I also reproduced it end-to-end through the real AgentLoop (/tmp/test_e2e_iso.py), where a Telegram turn's system prompt contained the WeChat summary.

    2. If an error was reported in the issue description, I have confirmed that my fix resolves the error. Yes. After the change: entries carry session_key; a Telegram session's prompt contains only its own entries, WeChat/cron/legacy entries are excluded; in unified_session mode user-channel entries are shared while cron:/dream:/heartbeat entries stay isolated (unless the current session is that internal session). Verified both at MemoryStore/ContextBuilder level and through a real AgentLoop._process_message turn.

    3. I have thought about edge cases and made sure my fix handles them as well.

    • Every archive path propagates the session key: LLM summary (Consolidator.archive), raw fallback inside archive(), replay-overflow consolidation, token-budget consolidation rounds, idle compaction (compact_idle_session), file-cap raw_archive in both loop paths, and /new archival.
    • Legacy entries (no session_key), non-string/empty keys, and empty current session key are handled (never injected).
    • session_key is only written when a non-empty string, so the JSONL stays backward compatible; readers ignore unknown/missing fields; cursor-corruption handling unchanged.
    • Dream (build_dream_prompt) still consumes all entries — global digestion is intentional and unchanged.
    • /new still clears the session before scheduling the background archive (never waits for archival success), and close_mcp() drains background tasks with a bounded loop plus an idempotent, ValueError-safe done-callback (the old list.remove callback could raise after clear()).
    • Token-estimation probe now passes the session key so estimation matches the real prompt.
    • Docs (docs/memory.md, memory SKILL.md) updated for the new record shape.

    4. I have run existing tests in the repository that might have been affected and confirmed that they pass. Ran the full suite. Non-agent tests: only one pre-existing failure (tests/webui/test_mcp_presets_api.py::test_test_mcp_preset_connects_and_reports_tools, which also fails on the pristine checkout). In tests/agent, 1094 pass and 9 fail: 5 use hand-written stubs whose signatures predate the required keyword-only session_key (archive(_messages)), and 4 assert the old, now-explicitly-forbidden behavior of injecting legacy entries without a session key. Both categories are the intended behavior change described in the evaluation contract; I did not modify any test files.

Responses were produced in different, uncontrolled contexts (system prompts, settings, prior turns). This is not a benchmark.