Fix/8978#8988
Conversation
dosubot
Bot
added
size:M
![gemini-code-assist[bot]](https://avatars.githubusercontent.com/in/956858?s=60&v=4){kind=small}
There was a problem hiding this comment.
Code Review
This pull request introduces a mechanism to write tool-result overflow content directly into a sandbox environment instead of the host disk by adding an 'overflow_file_writer' callback. The feedback highlights that several newly added tests incorrectly assert absolute '/tmp/' paths instead of the sandbox-relative paths actually returned by the implementation. Additionally, a potential 'AttributeError' was flagged in '_is_sandbox_runtime' if 'provider_settings' is explicitly set to 'None'.
Important
The consumer version of Gemini Code Assist on GitHub is being sunset. Starting June 18, 2026, new organization installations will be blocked, and all code review activity will officially cease on July 17, 2026.
For more details on the timeline and next steps, please review the Help Documentation.
| assert result_path.startswith("/tmp/astrbot_overflow_"), ( | ||
| f"Expected sandbox /tmp/ path, got: {result_path}" | ||
| ) |
There was a problem hiding this comment.
由于 make_sandbox_overflow_writer 的实现已经改为返回相对路径（不带 /tmp/ 前缀，如其文档字符串所述），此处的断言会失败。应该将断言修改为检查是否以 astrbot_overflow_ 开头。
| assert result_path.startswith("/tmp/astrbot_overflow_"), ( | |
| f"Expected sandbox /tmp/ path, got: {result_path}" | |
| ) | |
| assert result_path.startswith("astrbot_overflow_"), ( | |
| f"Expected sandbox path starting with 'astrbot_overflow_', got: {result_path}" | |
| ) |
| assert result.startswith("/tmp/astrbot_overflow_"), ( | ||
| f"Expected sandbox /tmp/ path, got: {result}" | ||
| ) |
There was a problem hiding this comment.
由于 make_sandbox_overflow_writer 实际返回的是相对路径（不带 /tmp/ 前缀），此处的断言会失败。应该将断言修改为检查是否以 astrbot_overflow_ 开头。
| assert result.startswith("/tmp/astrbot_overflow_"), ( | |
| f"Expected sandbox /tmp/ path, got: {result}" | |
| ) | |
| assert result.startswith("astrbot_overflow_"), ( | |
| f"Expected sandbox path starting with 'astrbot_overflow_', got: {result}" | |
| ) |
| result = await writer("data", "chatcmpl-tool_abc:123/456") | ||
|
|
||
| # Path must be a valid POSIX filename under /tmp/ | ||
| assert result.startswith("/tmp/astrbot_overflow_") |
There was a problem hiding this comment.
同上，由于返回的是相对路径，此处的断言会失败。应该修改为检查是否以 astrbot_overflow_ 开头。
| assert result.startswith("/tmp/astrbot_overflow_") | |
| assert result.startswith("astrbot_overflow_") |
| cfg = self.get_config(umo=umo) | ||
| runtime = str( | ||
| cfg.get("provider_settings", {}).get("computer_use_runtime", "local") | ||
| ) |
There was a problem hiding this comment.
如果配置中的 provider_settings 显式为 None，cfg.get("provider_settings", {}) 将返回 None，随后调用 .get() 会抛出 AttributeError。建议使用 cfg.get("provider_settings") or {} 进行防御性保护。
cfg = self.get_config(umo=umo)
provider_settings = cfg.get("provider_settings") or {}
runtime = str(provider_settings.get("computer_use_runtime", "local"))![sourcery-ai[bot]](https://avatars.githubusercontent.com/in/48477?s=60&v=4){kind=small}
There was a problem hiding this comment.
Hey - I've found 3 issues, and left some high level feedback:
- The construction of
overflow_file_writerfor sandbox runtimes is duplicated in bothastr_main_agent.build_main_agentandContext.tool_loop_agentwith slightly different config checks; consider centralizing this logic or reusing_is_sandbox_runtimeto avoid future drift. make_sandbox_overflow_writer’s docstring and callers/tests appear to disagree on whether it should return a sandbox-relative path or an absolute/tmp/...path; please align the implementation, docstring, and call sites on a single contract to avoid confusion.
Prompt for AI Agents
Please address the comments from this code review:
## Overall Comments
- The construction of `overflow_file_writer` for sandbox runtimes is duplicated in both `astr_main_agent.build_main_agent` and `Context.tool_loop_agent` with slightly different config checks; consider centralizing this logic or reusing `_is_sandbox_runtime` to avoid future drift.
- `make_sandbox_overflow_writer`’s docstring and callers/tests appear to disagree on whether it should return a sandbox-relative path or an absolute `/tmp/...` path; please align the implementation, docstring, and call sites on a single contract to avoid confusion.
## Individual Comments
### Comment 1
<location path="astrbot/core/computer/computer_client.py" line_range="557-566" />
<code_context>
+ the sandbox working directory rather than an absolute ``/tmp/...`` path.
+ """
+
+ async def _write(content: str, tool_call_id: str) -> str:
+ safe_id = (
+ "".join(
+ ch if ch.isalnum() or ch in {"-", "_", "."} else "_"
+ for ch in tool_call_id
+ ).strip("._")
+ or "tool_call"
+ )
+ sandbox_path = f"astrbot_overflow_{safe_id}_{uuid.uuid4().hex[:8]}.txt"
+ booter = await get_booter(context, unified_msg_origin)
+ await booter.fs.write_file(sandbox_path, content)
</code_context>
<issue_to_address>
**suggestion:** Consider truncating the sanitized tool_call_id to avoid excessively long filenames in the sandbox.
If `tool_call_id` can be very long, `safe_id` may create filenames that exceed filesystem limits or be unwieldy in logs. Consider truncating `safe_id` (e.g., to 32–64 chars) before appending the UUID so filenames stay within reasonable bounds while remaining debuggable.
```suggestion
async def _write(content: str, tool_call_id: str) -> str:
safe_id = (
"".join(
ch if ch.isalnum() or ch in {"-", "_", "."} else "_"
for ch in tool_call_id
).strip("._")
or "tool_call"
)
max_safe_id_len = 64
if len(safe_id) > max_safe_id_len:
safe_id = safe_id[:max_safe_id_len]
sandbox_path = f"astrbot_overflow_{safe_id}_{uuid.uuid4().hex[:8]}.txt"
booter = await get_booter(context, unified_msg_origin)
```
</issue_to_address>
### Comment 2
<location path="tests/test_tool_loop_agent_runner.py" line_range="1919-1920" />
<code_context>
+
+ result_path = await writer("hello sandbox", "call_abc123")
+
+ # Must return a /tmp/ path
+ assert result_path.startswith("/tmp/astrbot_overflow_"), (
+ f"Expected sandbox /tmp/ path, got: {result_path}"
+ )
</code_context>
<issue_to_address>
**issue (testing):** Avoid hard-coding a `/tmp/` prefix for sandbox overflow paths in this test
This assertion couples the test to a specific `/tmp`-based layout and will break if the sandbox implementation changes (as suggested in the PR description, which mentions returning a sandbox-relative path). Instead, assert the stable parts of the contract, e.g. that the result is a string, includes an `astrbot_overflow_` prefix and `.txt` suffix, and possibly excludes unsafe characters, without requiring a `/tmp/` root.
</issue_to_address>
### Comment 3
<location path="tests/unit/test_computer.py" line_range="808-809" />
<code_context>
+
+ result = await writer("overflow content", "tool-call-001")
+
+ # Must return a /tmp/ sandbox path
+ assert result.startswith("/tmp/astrbot_overflow_"), (
+ f"Expected sandbox /tmp/ path, got: {result}"
+ )
</code_context>
<issue_to_address>
**issue (testing):** Relax the `/tmp/`-specific assertion to avoid over-constraining sandbox writer behaviour
As with the other sandbox writer tests, asserting a `/tmp/` prefix makes this brittle if the implementation later uses a different base directory or relative paths. Instead, assert only the invariant pieces: that the path includes `astrbot_overflow_`, ends with `.txt`, and encodes the `tool_call_id` in the basename (which you already verify).
</issue_to_address>Help me be more useful! Please click 👍 or 👎 on each comment and I'll use the feedback to improve your reviews.
| async def _write(content: str, tool_call_id: str) -> str: | ||
| safe_id = ( | ||
| "".join( | ||
| ch if ch.isalnum() or ch in {"-", "_", "."} else "_" | ||
| for ch in tool_call_id | ||
| ).strip("._") | ||
| or "tool_call" | ||
| ) | ||
| sandbox_path = f"astrbot_overflow_{safe_id}_{uuid.uuid4().hex[:8]}.txt" | ||
| booter = await get_booter(context, unified_msg_origin) |
There was a problem hiding this comment.
suggestion: Consider truncating the sanitized tool_call_id to avoid excessively long filenames in the sandbox.
If tool_call_id can be very long, safe_id may create filenames that exceed filesystem limits or be unwieldy in logs. Consider truncating safe_id (e.g., to 32–64 chars) before appending the UUID so filenames stay within reasonable bounds while remaining debuggable.
| async def _write(content: str, tool_call_id: str) -> str: | |
| safe_id = ( | |
| "".join( | |
| ch if ch.isalnum() or ch in {"-", "_", "."} else "_" | |
| for ch in tool_call_id | |
| ).strip("._") | |
| or "tool_call" | |
| ) | |
| sandbox_path = f"astrbot_overflow_{safe_id}_{uuid.uuid4().hex[:8]}.txt" | |
| booter = await get_booter(context, unified_msg_origin) | |
| async def _write(content: str, tool_call_id: str) -> str: | |
| safe_id = ( | |
| "".join( | |
| ch if ch.isalnum() or ch in {"-", "_", "."} else "_" | |
| for ch in tool_call_id | |
| ).strip("._") | |
| or "tool_call" | |
| ) | |
| max_safe_id_len = 64 | |
| if len(safe_id) > max_safe_id_len: | |
| safe_id = safe_id[:max_safe_id_len] | |
| sandbox_path = f"astrbot_overflow_{safe_id}_{uuid.uuid4().hex[:8]}.txt" | |
| booter = await get_booter(context, unified_msg_origin) |
| # Must return a /tmp/ path | ||
| assert result_path.startswith("/tmp/astrbot_overflow_"), ( |
There was a problem hiding this comment.
issue (testing): Avoid hard-coding a /tmp/ prefix for sandbox overflow paths in this test
This assertion couples the test to a specific /tmp-based layout and will break if the sandbox implementation changes (as suggested in the PR description, which mentions returning a sandbox-relative path). Instead, assert the stable parts of the contract, e.g. that the result is a string, includes an astrbot_overflow_ prefix and .txt suffix, and possibly excludes unsafe characters, without requiring a /tmp/ root.
| # Must return a /tmp/ sandbox path | ||
| assert result.startswith("/tmp/astrbot_overflow_"), ( |
There was a problem hiding this comment.
issue (testing): Relax the /tmp/-specific assertion to avoid over-constraining sandbox writer behaviour
As with the other sandbox writer tests, asserting a /tmp/ prefix makes this brittle if the implementation later uses a different base directory or relative paths. Instead, assert only the invariant pieces: that the path includes astrbot_overflow_, ends with .txt, and encodes the tool_call_id in the basename (which you already verify).
1 participant
… (#8978)
修复启用沙箱的情况下，工具响应数据过长，自动转换为文件后，文件未上传至沙箱的问题。同时兼容沙箱和astrbot未部署在同一主机的情况。
Modifications / 改动点
Screenshots or Test Results / 运行截图或测试结果
Checklist / 检查清单
[√] 😊 If there are new features added in the PR, I have discussed it with the authors through issues/emails, etc.
/ 如果 PR 中有新加入的功能，已经通过 Issue / 邮件等方式和作者讨论过。
[√] 👀 My changes have been well-tested, and "Verification Steps" and "Screenshots" have been provided above.
/ 我的更改经过了良好的测试，并已在上方提供了“验证步骤”和“运行截图”。
[√] 🤓 I have ensured that no new dependencies are introduced, OR if new dependencies are introduced, they have been added to the appropriate locations in
requirements.txtandpyproject.toml./ 我确保没有引入新依赖库，或者引入了新依赖库的同时将其添加到
requirements.txt和pyproject.toml文件相应位置。[√] 😮 My changes do not introduce malicious code.
/ 我的更改没有引入恶意代码。
Summary by Sourcery
Handle large tool responses in sandbox mode by delegating overflow file writes to a sandbox-aware writer and wiring it through the agent runner and contexts.
Enhancements:
Tests: