"Why does assistant.threads.setStatus fail?", "should this DM have assistant status/title?", "does Chat tab DM count as an assistant thread?"
Read Slack docs for assistant_thread_started, assistant_thread_context_changed, message.im, and assistant.threads.* first, then load ${CLAUDE_SKILL_ROOT}/references/assistant-thread-apis.md and ${CLAUDE_SKILL_ROOT}/references/chat-sdk-payload-contract.md
Long-running behavior
"No feedback while it runs", "show progress", "stream output"
If the request is ambiguous, ask one focused question and continue after clarification.
Step 2: Load only relevant references and implement
Use the selected reference files as the implementation guide. Keep SKILL.md high-level and put details in references.
Slack assistant-thread guardrails:
Use Slack's current inbound event payload as the source of truth for assistant-thread API calls. For non-DM message events, use the live channel plus thread_ts ?? ts. For message.im, require the live channel and explicit thread_ts. For lifecycle events, use assistant_thread.channel_id and assistant_thread.thread_ts.
Do not invent assistant-thread identifiers from persisted state unless Slack's docs explicitly require it.
Separate reply continuity from assistant-thread API eligibility. A stored root timestamp can be valid for reply threading without being valid for assistant.threads.*.
Treat assistant_thread_started and assistant_thread_context_changed differently. Context changes can refresh prompts/context, but should not clobber a conversation-specific thread title back to a generic default.
Conversation-specific thread titles should come from the earliest human message the runtime actually knows about for that thread, using the lightweight title model. Do not base titles on assistant reply text or a later follow-up.
Title generation may run in parallel with the main assistant turn, but it must not delay assistant reply generation or visible reply delivery.
Assistant status is best effort. Do not make Slack status writes part of the critical path for tool/model execution.
If debugging a live repro through the example app, verify whether the app is executing packages/junior/dist/* output before trusting source edits against runtime behavior.
Step 3: Enforce project conventions
When modifying this repository:
Keep tool behavior aligned with AI Gateway primitives already in use.
Avoid reintroducing deprecated custom search integrations when Gateway-native tools exist.
Preserve webhook waitUntil behavior so long-running handlers finish after HTTP response.
Assistant threads: assistant.threads.* calls use the live inbound assistant-thread context; non-DM message events may use thread_ts ?? ts, message.im must carry explicit thread_ts, and runtime code does not synthesize DM assistant roots for status/title updates.
Accessibility: block messages include an adequate top-level fallback text strategy.
Latency UX: user sees immediate feedback for long-running tasks.
Streaming/progress: behavior is observable during tool/model execution, not only at completion.
Failure mode: errors return actionable responses rather than silent stalls.
