Module distill 

Distillation: builds the prompt with evidence (git + sessions) and calls the LLM via a configurable command (default “claude -p”). Injectable command means tests use cat and users can swap LLMs without changing code.

Enums

Confidence

How well AI sessions align with git evidence for a branch.

Functions

build_prompt

Builds the context-reconstruction prompt for an open loop.

compute_confidence

Derives a confidence level from matched session excerpts.

format_dry_run

Shows git and session evidence that would feed distillation, without calling the LLM.

run_llm

Runs the LLM command with the prompt on stdin and returns stdout.

with_sources

Appends the ## Sources section to the LLM-generated document.