Allow users to configure which OpenClaw agent the integration communicates
with, instead of always defaulting to the implicit 'main' agent.
Changes:
- const.py: add CONF_AGENT_ID, DEFAULT_AGENT_ID='main', ATTR_AGENT_ID
- api.py: accept agent_id in constructor; _headers() now includes the
x-openclaw-agent-id header on every request; async_send_message and
async_stream_message accept an optional per-call agent_id override
- config_flow.py: expose agent_id as a text field in both the manual
setup step and the options flow (Settings → Integrations → Configure)
- __init__.py: read agent_id from options/data and pass it to the API
client; add optional agent_id field to the send_message service schema
so automations can address a specific agent per-call
- services.yaml: document the new agent_id field on send_message
- strings.json / translations/en.json: add UI labels for the new option
The gateway routing header x-openclaw-agent-id is always sent; when
agent_id is 'main' (the default) the gateway behaviour is unchanged.
A per-call override on send_message takes precedence over the config.
Implemented with assistance from an AI coding agent (BMO) with human
review. Bot-assisted contribution reviewed by a human before submission.
- Implemented HTTPS/SSL support for OpenClaw gateways in `lan_https` mode.
- Added configuration options for SSL certificate verification.
- Improved auto-discovery for `lan_https` access mode.
- Updated documentation in README and changelog for new features and fixes.
- Fixed connection errors related to SSL certificate verification.
- Documented changes in changelog for version 0.1.3.
- Improved error handling in async_check_connection to propagate OpenClawApiError.
- Added clear error messages for users regarding OpenAI API settings in config flow.
- Updated translations to include new error messages related to OpenAI API status.
BMO (OpenClaw)