fix: apply max_agent_step config to cron and background task agents

[Shaofei215]

Fixes #9800
修复了 #9800

Modifications / 改动点

• Read provider_settings.max_agent_step in the cron (future task) wake-up path (astrbot/core/cron/manager.py) and the background task result wake-up path (astrbot/core/astr_agent_tool_exec.py), and pass it to runner.step_until_done() instead of the hardcoded 30.
  在 Cron 未来任务唤醒路径（astrbot/core/cron/manager.py）与后台任务结果唤醒路径（astrbot/core/astr_agent_tool_exec.py）中读取 provider_settings.max_agent_step，传入 runner.step_until_done()，替换写死的 30。

• Parse the value with the existing coerce_int_config utility: bool / non-numeric string / None falls back to the default 30 with a warning. This is consistent with the main pipeline's [Bug] 更新4.0前如果更改过唤醒词，发出的任何请求不会有回应 #2622 bool workaround and the existing usage in the deerflow runner / request retry logic. Behavior is unchanged for users who have not customized this setting.
  使用既有的 coerce_int_config 工具解析取值：布尔 / 非数字字符串 / None 回退到默认值 30 并输出警告。与主管线的 [Bug] 更新4.0前如果更改过唤醒词，发出的任何请求不会有回应 #2622 布尔 workaround 及 deerflow runner / 请求重试逻辑的既有用法保持一致。未自定义该项配置的用户行为不变。

• Add 8 parametrized unit tests (4 cases × 2 paths) covering: configured value propagation, missing-key fallback, boolean pollution fallback, and numeric string coercion.
  新增 8 个参数化单元测试（4 种用例 × 2 条路径），覆盖：配置值传播、缺省回退、布尔污染回退、数字字符串转换。

• This is NOT a breaking change. / 这不是一个破坏性变更。

Screenshots or Test Results / 运行截图或测试结果

Lint and unit tests (all executed locally on the PR branch) / 代码检查与单元测试（均在 PR 分支本地执行）：

$ ruff check .
All checks passed!

$ ruff format --check astrbot/core/cron/manager.py astrbot/core/astr_agent_tool_exec.py tests/unit/test_cron_manager.py tests/unit/test_astr_agent_tool_exec.py
4 files already formatted

$ pytest tests/unit/test_cron_manager.py tests/unit/test_astr_agent_tool_exec.py -q
57 passed in 6.66s

Regression guard verification: with the fix reverted (git checkout master -- astrbot/core/cron/manager.py astrbot/core/astr_agent_tool_exec.py), the new value-propagation test cases fail as expected (4 failed, 4 passed), proving the tests actually protect against the regression.
回归防护验证：将修复还原后（git checkout master -- astrbot/core/cron/manager.py astrbot/core/astr_agent_tool_exec.py），新增的取值传播测试用例如预期失败（4 failed, 4 passed），证明测试确实能防护该回归。

---

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。
  (Bug fix only, no new features — the fix targets the behavior reported in [Bug] 未来任务被写死只能调用30次工具 #9800. / 仅为缺陷修复，无新增功能，修复对象即 [Bug] 未来任务被写死只能调用30次工具 #9800 所报告的行为。)
• 👀 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.txt and pyproject.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

Honor max_agent_step configuration for cron and background-task agent execution while retaining safe fallback behavior.

Bug Fixes:

• Apply the configured provider_settings.max_agent_step limit to cron-triggered and background-task agent wake-ups instead of always using the default limit.

Enhancements:

• Normalize invalid or missing step-limit configuration through the existing integer configuration handling, preserving the default and minimum safeguards.

Tests:

• Add parameterized unit coverage for configured values, defaults, boolean and numeric-string inputs, and minimum-value clamping across both wake-up paths.

[sourcery-ai]

Hey - I've found 1 issue

Prompt for AI Agents

Please address the comments from this code review:

## Individual Comments

### Comment 1
<location path="astrbot/core/astr_agent_tool_exec.py" line_range="556" />
<code_context>
         cron_event.role = event.role
         cfg = ctx.get_config(umo=event.unified_msg_origin) or {}
         provider_settings = cfg.get("provider_settings") or {}
+        agent_max_step = coerce_int_config(
+            provider_settings.get("max_agent_step", 30),
+            default=30,
+            field_name="provider_settings.max_agent_step",
+        )
         config = MainAgentBuildConfig(
</code_context>
<issue_to_address>
**issue (bug_risk):** Non-positive `max_agent_step` values are accepted because these calls omit `min_value`. The local runner skips all normal steps and immediately forces a final response for zero or negative values, while the DeerFlow runner raises `ValueError`, so an invalid configured value breaks or truncates both wake-up paths instead of falling back to a usable default.

**Triggers:** When `provider_settings.max_agent_step` is configured as zero or a negative integer.

**Suggested fix:** Pass `min_value=1` to `coerce_int_config` so non-positive values fall back to the default or minimum instead of reaching `step_until_done`.

```suggestion
            min_value=1,
            field_name="provider_settings.max_agent_step",
```
</issue_to_address>

Sourcery assessment

Needs a human reviewer. 1 finding to address first, and if the configured limit is wrong, cron or background agents may perform extra model/tool steps, increasing compute and potentially triggering additional tool side effects before the run stops. Reverting restores the fixed 30-step behavior, but any messages or other actions already taken during an overlong run would not be undone.

Blocking findings: astrbot/core/astr_agent_tool_exec.py:556

---

Sourcery is free for open source - if you like our reviews please consider sharing them ✨

• X
• Mastodon
• LinkedIn
• Facebook

_{Help me be more useful! Please click 👍 or 👎 on each comment and I'll use the feedback to improve your reviews.}