slack-search

Lark IM Search

This skill provides guidance for effectively finding information in Lark IM — messages, files, group chats, and people. The Lark connector supersedes Slack here: route all reads through the curated lark_* tools. For search grammar and pagination depth, delegate to the lark-im skill.

Depth core: ../../connectors/LARK-PATTERNS.md (P1 identity, P3 token economy — project reads with jq), ../../connectors/LARK-RECIPES.md, ../../connectors/LARK-FUSION.md.

When to Use

Apply this skill whenever you need to find information in Lark IM — locating messages, conversations, files, or people, or gathering context before answering a question about what's happening in chat.

Search Tools Overview

Goal	Tool	Notes
Find messages / conversations	lark_im_search	Search chat history by keyword. Project results with jq (P3) — pull only the fields you need.
Find a person (name/email/role → open_id)	lark_contact_search	P1 — resolve identity before any follow-up send/assign.
Find files shared in chat	lark_im_search	Filter to file/attachment content; see the lark-im skill for content-type filters.

Anything without a curated tool goes through lark_api (Lark OpenAPI passthrough). For advanced filters, pagination, and reading full thread context, defer to the lark-im skill.

Search Strategy

Start Broad, Then Narrow

1. Begin with a simple keyword or natural language question.
2. If too many results, add filters (group chat, sender, date range).
3. If too few results, remove filters and try synonyms or related terms.

Choose the Right Approach

• Natural language questions (e.g., "What is the deadline for project X?") — best for fuzzy, conceptual searches where you don't know exact keywords.
• Keyword search (e.g., project X deadline) — best for finding specific, exact content.

Use Multiple Searches

Don't rely on a single search. Break complex questions into smaller searches:

• Search for the topic first
• Then resolve and search for specific people's contributions (lark_contact_search → filter by sender)
• Then narrow to specific group chats

Token Economy (P3)

Lark IM search can return large payloads. Always pass a jq projection to lark_im_search so you keep only the fields you need (sender, timestamp, snippet) instead of full message bodies. Pull the full payload only when you'll actually use it.

Resolving People in Results

When a result references a person by open_id, resolve it to a name/department with lark_contact_search (P1). Likewise, before any follow-up — a reply, a mention, an assignment — resolve the target's open_id first; IDs cannot be guessed.

Following Up on Results

After finding relevant messages:

• Resolve any open_ids to people via lark_contact_search.
• To reply or surface a digest, hand off to the slack-messaging skill (lark_im_send / lark_im_card_send).
• To turn a found decision/action item into durable state, log it to a Lark Base (P5, lark_base_record_upsert) or create a task (lark_task_create).

Common Pitfalls

• Search is not real-time. Very recent messages may not appear in search results yet.
• Don't guess IDs. Always resolve a name/email to an open_id via lark_contact_search before acting on a result.
• Don't over-fetch. Without a jq projection a broad search can burn tens of thousands of tokens — project first (P3).