fix: pass persona skills to subagents

[he-yufeng]

Summary

• carry persona skill selections into subagent handoff definitions
• append the selected skill prompt when running the subagent handoff
• keep inline subagents without selected skills unchanged

Addresses the subagent skill-loading part of #8099. The duplicate-message behavior reported there is separate and not changed here.

To verify

• python -m py_compile astrbot\core\agent\agent.py astrbot\core\subagent_orchestrator.py astrbot\core\astr_agent_tool_exec.py tests\unit\test_subagent_orchestrator.py tests\unit\test_astr_agent_tool_exec.py
• python -m pytest tests\unit\test_subagent_orchestrator.py tests\unit\test_astr_agent_tool_exec.py::test_build_handoff_skills_prompt_filters_selected_skills tests\unit\test_astr_agent_tool_exec.py::test_execute_handoff_appends_agent_skills_prompt -q
• python -m ruff check astrbot\core\agent\agent.py astrbot\core\subagent_orchestrator.py astrbot\core\astr_agent_tool_exec.py tests\unit\test_subagent_orchestrator.py tests\unit\test_astr_agent_tool_exec.py
• git diff --check

Summary by Sourcery

Propagate persona-defined skills to subagent handoffs and include the corresponding skills prompt in subagent executions.

New Features:

• Support associating skills with agents and subagents via the Agent model and subagent orchestrator configuration.
• Generate and append a skills-based system prompt segment when executing subagent handoffs, based on the subagent's selected skills.

Bug Fixes:

• Ensure subagent persona skills are correctly carried through from persona configuration into handoff agent definitions.

Tests:

• Add unit tests covering subagent skill normalization from config, skills propagation into handoff agents, and skills prompt construction and usage during handoff execution.

[sourcery-ai]

Hey - I've left some high level feedback:

• The Agent.skills field is typed as list[str] | None but uses field(default_factory=list), which means it will never be None by default; consider either changing the type to list[str] or using a None default to better distinguish between None and an empty list.
• In _build_handoff_system_prompt, concatenating instructions and the skills prompt with fixed "\n" delimiters can easily produce extra leading/trailing blank lines; consider stripping the inputs or joining non-empty segments to normalize the final system prompt formatting.

Prompt for AI Agents

Please address the comments from this code review:

## Overall Comments
- The `Agent.skills` field is typed as `list[str] | None` but uses `field(default_factory=list)`, which means it will never be `None` by default; consider either changing the type to `list[str]` or using a `None` default to better distinguish between `None` and an empty list.
- In `_build_handoff_system_prompt`, concatenating `instructions` and the skills prompt with fixed `"\n"` delimiters can easily produce extra leading/trailing blank lines; consider stripping the inputs or joining non-empty segments to normalize the final system prompt formatting.

---

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.}

[gemini-code-assist]

Code Review

This pull request introduces a 'skills' system for sub-agents, allowing them to have specific capabilities beyond basic tools. It updates the Agent model, the handoff execution logic to include skill-based prompts in the system message, and the orchestrator to handle skill configuration. Feedback suggests improving the system prompt construction to avoid formatting issues and refining the skill list normalization to prevent None values from being converted to strings.

[gemini-code-assist]

The construction of the system prompt can lead to leading or trailing newlines if instructions is empty or skills_prompt is present. It's generally cleaner to join the prompt parts with double newlines to clearly separate sections for the LLM.

        prompt_parts = []
        if instructions:
            prompt_parts.append(instructions.strip())

        skills_prompt = cls._build_handoff_skills_prompt(skill_names, runtime)
        if skills_prompt:
            prompt_parts.append(skills_prompt.strip())

        return "\n\n".join(prompt_parts)

[gemini-code-assist]

The list comprehension str(s).strip() will convert None values in the skills list to the string "None". It's safer to filter out non-truthy values before converting to string.

Suggested change

	skills = [str(s).strip() for s in skills if str(s).strip()]
	skills = [str(s).strip() for s in skills if s and str(s).strip()]