Consume Lark/Feishu real-time events as NDJSON streams via lark-cli.
npx skills add https://github.com/CalWade/aigc-creator --skill lark-event-calwade
Please help me install this Agent Skill. Skill: lark-event Source: https://github.com/CalWade/aigc-creator/tree/main/.agents/skills/lark-event Command: npx skills add https://github.com/CalWade/aigc-creator --skill lark-event-calwade
lark-cli event consume, with a subprocess contract designed for agent callers. ## Core Features & Use Cases - Event discovery and schema inspection: List all subscribable EventKeys with lark-cli event list and inspect each key's params and output schema with lark-cli event schema before writing jq filters. - Bounded or unbounded consumption: Stream events to stdout as NDJSON, with --max-events / --timeout for bounded runs, --jq for filtering and projection, and a stderr ready-marker plus structured exit codes for reliable subprocess orchestration. - Domain references: Detailed catalogs for 11 IM EventKeys, 2 VC EventKeys, 1 Minutes EventKey, and 1 Whiteboard EventKey, including field semantics, auth scopes, and jq recipes. - Use Case: Build a Lark bot that watches im.message.receive_v1, filters for group text messages mentioning a keyword, and pipes matching events into a downstream summarization pipeline. ## Quick Start Ask the AI to listen for incoming Lark messages by running lark-cli event consume im.message.receive_v1 as the bot identity and show you the streamed events.High-intent search queries and answers about installing and using this skill.
Run lark-cli event consume im.message.receive_v1 --as bot to stream every incoming message as NDJSON to stdout. Add --jq filters to narrow by chat type, message type, or sender, and use --max-events or --timeout for bounded runs.
Pass a jq expression via the --jq flag, for example --jq 'select(.message_type=="text") | .content'. Check lark-cli event schema <key> --json first to learn the jq_root_path and field types, since some fields are pre-decoded and must not be passed through fromjson.
No, each event consume process handles exactly one EventKey. To listen to multiple keys, run one subprocess per key; they all share a single local bus daemon, so overhead is small and each stream keeps one consistent output shape.
For unbounded runs, closing stdin (for example via /dev/null or nohup defaults) signals a graceful shutdown. Keep stdin open with a never-ending source like tail -f /dev/null, or run bounded with --max-events or --timeout, which ignore stdin EOF.
kill -9 skips the OAPI unsubscribe step for EventKeys with PreConsume hooks, leaking server-side subscriptions and causing 'subscription already exists' errors or duplicate delivery later. Use SIGTERM or close stdin instead for graceful cleanup.
Yes, the --as user|bot|auto flag selects the identity, but support varies by EventKey. For example, VC and Minutes keys require --as user, while board.whiteboard.updated_v1 supports both user and bot with manage access to the whiteboard.
loongclaw-aiCreate, fetch, update, and search Feishu documents via lark-cli shortcuts.
wuyyyyyouStreams new Gmail messages as NDJSON output via configured sink.
yokingmaResolve Lark/Feishu contact names or emails into open_id values.
openaiDesign, secure, and operate Twilio webhook endpoints for messaging, voice, and event streams.