Skip to content

nanocoai/nanoclaw / .claude/skills/slack-agent-flow/container/skills/slack-construct-agents

slack-construct-agents

Verified

Standing context for Slack sibling agents — one room per team, creator-posted introductions, bot-to-bot hop discipline. Ships as instructions.md composed into the group's CLAUDE.md at spawn; there is no workflow to invoke.

FileKindSize
SKILL.mdskill md1.2 KB
instructions.mdmarkdown1.8 KB

instructions.md

View on GitHub
## Your Slack sibling agents

You may be part of a construct of agents on Slack: sibling agents (each with its own bot,
DM, and workspace) plus shared rooms where humans, you, and siblings talk. Standing rules
for the sibling half:

- **Teams get ONE room.** When the user asks for several agents for one project ("build me
  a team"), create each agent with `create_agent({ ..., room: 'none' })` — they still get
  their operator DM — then open a single shared room with
  `create_room({ name, purpose, agents: [all of them] })`. Never let each create open its
  own room: that yields N separate three-way rooms nobody wants. `add_to_room` works for
  later growth, but Slack group DMs never grow in place — the room MOVES to a new
  conversation (everyone is re-wired automatically; the old one keeps working), so prefer
  creating rooms complete.
- **You introduce agents you create.** When a shared room comes with an agent you created,
  YOU post the introduction in the room — the host posts nothing there. You'll get a
  system nudge telling you which destination to use. Keep it to 1-2 lines in your own
  voice: say what the new agent is for and tag them with their `<@bot-user-id>` mention
  (send it literally; it renders as a mention). No mechanics, no member lists — the room's
  canvas tab already holds that.
- **Bot-to-bot hop budget.** The platform may cap consecutive bot-to-bot messages (~6)
  until a human speaks again, but do not rely on it — self-limit. Don't ping-pong with
  siblings: do the work, converge, hand back to the human.
- **Persist durable facts.** Conversations are per-session; rooms and DMs don't share
  history. Anything worth keeping (decisions, preferences, ongoing state) goes in your
  memory directory, not just the chat transcript.
reverse-engineer-anythingmorluto/reaReverse engineer native, managed, Electron/JavaScript, packaged, firmware, and browser targets with REA, and build agent environments (AgentEnv bundles) of web applications from a live instance or a capture, and of Linux desktop apps with a task from a request. Use shipped-artifact or requested runtime evidence to explain features, compare versions, decompile code, or guide a reconstruction. Skip REA for ordinary source-repository architecture analysis.Stars: 471kUpdated Claude CodeCodexVerifiedInstall
diagnosing-superpowersobra/superpowersUse when a superpowers session went wrong and your human partner wants to know why — repeated work, ignored plans, stumbles, poor results, a skill that didn't fire, "it took too long", "why is it so expensive", "what is it doing" — or wants to build a bug report for the superpowers maintainers, for the current session or a past one identified by id or path, on any harness.Stars: 322kUpdated Claude CodeCodexVerifiedInstall
dispatching-parallel-agentsobra/superpowersUse when facing 2+ independent tasks that can be worked on without shared state or sequential dependenciesStars: 322kUpdated Claude CodeCodexVerifiedInstall
executing-plansobra/superpowersUse when executing an implementation plan in the current session as the implementer yourself — your human partner chose inline execution, or no subagent tool is availableStars: 322kUpdated Claude CodeCodexVerifiedInstall