fix: correct grammar in guardrail, task output, and prompt templates

[Ricardo-M-L]

Summary

• Fix possessive it's → its in task.py output description
• Fix multiple grammar issues in llm_guardrail.py:
  • A feedback → Feedback (uncountable noun)
  • Wrong docstring: code generation → validation
  • a expert → an expert
  • Fix sentence fragments
• Fix broken grammar in en.json prompt templates:
  • must be the great and the most complete as possible → must be as great and complete as possible
  • must be outcome described → must be the outcome described

Test plan

• Verify LLM guardrail prompts read correctly
• Verify agent prompt templates render proper English

🤖 Generated with Claude Code

---

Note

Low Risk
Low risk: changes are limited to grammar/wording updates in model field descriptions, LLM guardrail prompts/docstrings, and English translation templates with no behavioral logic changes.

Overview
Polishes user-facing copy and prompt text across the codebase.

Fixes minor grammar in Task.output’s field description, improves wording in LLMGuardrail (feedback field description, docstrings, agent backstory, and validation prompt text), and corrects awkward English in en.json prompt templates used when tools aren’t available.

^{Reviewed by Cursor Bugbot for commit 0e59b34. Bugbot is set up for automated code reviews on this repo. Configure here.}