Configuration

WeaveRx reads settings from environment variables and CLI flags. CLI flags take precedence where both apply.

GitHub

Variable	Required	Description
GITHUB_TOKEN	For writes	Personal access token or GITHUB_TOKEN in Actions. Needed to post comments or apply labels. Optional for read-only triage on public repos.

LLM providers

WeaveRx uses LiteLLM. Set one provider's API key.

Variable	Provider	Description
XAI_API_KEY	Grok (default)	xAI API key
ANTHROPIC_API_KEY	Anthropic	Anthropic API key
OPENAI_API_KEY	OpenAI-compatible	OpenAI, Azure OpenAI, vLLM, or other OpenAI-compatible gateways
OPENAI_API_BASE	OpenAI-compatible	Custom API base URL (optional; omit for api.openai.com)
WEAVERX_LLM_PROVIDER	All	Override default provider: grok, anthropic, or openai
WEAVERX_LLM_MODEL	All	Override model id (e.g. anthropic/claude-3-5-haiku-20241022)

CLI equivalent: --llm-provider grok|anthropic|openai

Offline and mock modes

Variable	Default	Description
WEAVERX_MOCK	off	Set to 1, true, or yes for fully offline mock GitHub + mock LLM

CLI equivalents: --mock (offline GitHub + mock LLM), --mock-llm (real GitHub, mock LLM)

Safeguards

Local heuristic checks on generated drafts. No LLM calls on this path.

Variable	Default	Description
WEAVERX_SAFEGUARDS	1	Set to 0, false, or no to disable
WEAVERX_SAFEGUARD_ENTROPY_MAX	5.5	Shannon entropy threshold (bits per character)
WEAVERX_SAFEGUARD_MAX_CHARS	6000	Maximum draft length before flagging

CLI equivalents: --safeguards / --no-safeguards

GitHub Action

The composite action passes inputs as environment variables. See action.yml and GitHub Action in this site.

Quick reference

# Offline demo — no env vars
weaverx triage --repo Project-MONAI/MONAI --issue 42 --mock

# Real LLM, dry-run
export XAI_API_KEY=xai-...
weaverx triage --repo owner/repo --issue 1234 --dry-run

# Post comment (requires token + explicit confirm flags)
export GITHUB_TOKEN=ghp_...
weaverx triage --repo owner/repo --issue 1234 --post-comment --confirm

More examples: llm_provider_examples.md.