Skip to content

CLI reference

weaverx triage --repo owner/repo [--issue N | --recent N] [options]
Flag Description
--mock Offline mode (sample GitHub + mock LLM)
--mock-llm Real GitHub, offline mock LLM
--dry-run Analyze only; never write (default unless posting)
--json Machine-readable JSON output
--llm-provider grok, anthropic, or openai
--confirm Required to post comments or apply labels
--post-comment Post draft (needs --confirm)
--apply-labels Apply suggested labels (needs --confirm)
--privacy-insight Flag possible PHI/DICOM concerns (default: on)
--safeguards / --no-safeguards Local draft heuristics (default: on)
-v, --verbose Full draft, sources, debug logging

JSON output fields

Field Description
status ready_for_review or posted
issue GitHub issue metadata
analysis Raw LLM triage (category, priority, labels, raw draft, reasoning)
sources Issue excerpts that grounded the triage
draft_response Refined postable comment (use this for posting)
safeguard Local heuristic score, status, triggered flags, metrics
llm Provider and model used
duplicate_matches Similar recent issues (heuristic, not embeddings yet)

Full sample: sample_triage_output.json.