spacr.qt.ai.prompts

System prompts used by the AI console.

Two shapes:

• default_system_prompt() — spacr-aware assistant persona, used for the freeform chat panel.

• error_explainer_prompt() — prefix that turns a traceback into a concrete, numbered fix list capped at six steps.

Functions

default_system_prompt(→ str)	Return the spaCR-aware assistant persona used for freeform chat.
error_explainer_prompt(→ str)	Return the system prompt that turns a traceback into a numbered fix list.
tool_approval_note(→ str)	Tell the assistant how tool approval works in spaCR's console (GitHub #147).
wrap_error_for_prompt(→ str)	Turn a raw traceback into the user message body sent to the model.

Module Contents

spacr.qt.ai.prompts.default_system_prompt() → str

Return the spaCR-aware assistant persona used for freeform chat.

spacr.qt.ai.prompts.error_explainer_prompt() → str

Return the system prompt that turns a traceback into a numbered fix list.

spacr.qt.ai.prompts.tool_approval_note(provider_name: str) → str

Tell the assistant how tool approval works in spaCR’s console (GitHub #147).

The vendor CLIs run non-interactively, so they cannot show their own approval prompt. For Claude Code spaCR asks the user itself after the reply and resumes the session when the user allows the call; the other CLIs have no such route, so the assistant hands the command over instead.

Parameters:

provider_name – the provider’s short id, such as "claude".

Returns:

the paragraph to append to the system prompt.

spacr.qt.ai.prompts.wrap_error_for_prompt(traceback_text: str, active_app: str = '') → str

Turn a raw traceback into the user message body sent to the model.

Parameters:

traceback_text – the raw traceback text; surrounding whitespace is stripped and the rest is placed inside a fenced code block.