[Feature]: Support suppressing Agent's conversational response after script execution in Cron jobs

## Summary

Add an option to CoPaw Cron for "Agent" type tasks to suppress the Agent's textual reply after a script is successfully executed, allowing for a "silent" execution that only returns the script's output.

## Component(s) Affected

- [x] Core / Backend (app, agents, config, providers, utils, local_models)
- [x] Console (frontend web UI)
- [ ] Channels (DingTalk, Feishu, QQ, Discord, iMessage, etc.)
- [ ] Skills
- [ ] CLI
- [ ] Documentation (website)
- [ ] Tests
- [ ] CI/CD
- [ ] Scripts / Deploy

## Problem / Motivation

Currently, when a Cron task is configured as an "Agent" and instructed to run a specific script via its Prompt, the Agent almost always follows up with a conversational response (e.g., "I have executed the script for you...") after the tool call is finished.

In automated scheduling scenarios, users typically only care about the actual script output or the action performed. These extra responses consume unnecessary tokens and clutter notification logs or messaging channels (like Feishu or DingTalk) with redundant "filler" text.

## Proposed Solution

1. **Backend Logic**: Introduce a logic gate in the Cron execution flow. If a "Silent Mode" or "Skip Final Response" flag is set, the system should stop the process immediately after the tool/script returns its result and bypass the final LLM generation step.
2. **Console UI**: Add a toggle in the Cron Job configuration page (specifically when the type is set to "Agent"), such as `Silent Execution` or `Suppress Agent Reply`.

## Alternatives Considered

- **Prompt Engineering**: Adding instructions like "Do not speak, only execute the script" to the prompt. However, this is often unreliable across different models, as LLMs frequently ignore these constraints and provide a courtesy confirmation anyway.
- **Using "Skill" Type**: While Skills only execute the script, they lack the Agent's reasoning capabilities, such as pre-processing parameters or making conditional decisions based on the environment before running the script.

## Additional Context

This is particularly relevant for users who bridge CoPaw Cron to external notification channels. The goal is to receive clean data output rather than conversational filler from the Agent.