Slack logo

Slack

Organization
runxhq
slack

Read bounded Slack search, thread, link and file evidence through any compatible Slack binding, ready for an operator's triage and for message-send to reply.

Overview

Publisherrunxhq
Repositoryrunx
Skill nameslack
Stars
105
Forks
109
Bundled files
3
LicenseApache-2.0
Links
  • Markdown instructions

    A SKILL.md file the model loads on demand, so it only costs tokens when a request actually matches.

  • Works with any LLM

    AI skills are plain Markdown, not provider-specific code, so this works with GPT, Claude, Gemini, Grok, or a local model.

  • 3 bundled files

    Scripts, templates, and references the model can read while it works. Files are read-only and never executed.

  • Open source

    Published by runxhq on GitHub. Read the source before you install it.

Installation

Install the Slack AI skill in TypingMind to use it with any LLM, or drop it into another agent that reads SKILL.md.

1

Install in TypingMind

TypingMind installs a skill straight from its GitHub folder — it reads SKILL.md, bundles the resource files, and stores the result locally.

  1. Open the app and go to Plugins → Skills.
  2. Choose "Install from GitHub".
  3. Paste the skill folder URL below and confirm.
    https://github.com/runxhq/runx/tree/main/skills/slack
  4. Enable the skill in any chat where you want it available.
Plugins → Skills → Add skill → From GitHub URL, then paste the folder URL and press Continue.
2

Install in another agent

Any agent that reads the Agent Skills format can use this skill — copy the folder into that agent's skills directory.

Claude Code — .claude/skills
git clone --depth 1 https://github.com/runxhq/runx.git /tmp/runx
mkdir -p .claude/skills
cp -r /tmp/runx/skills/slack .claude/skills/slack
Restart Claude Code after copying so it picks up the new skill.

Use it in TypingMind

Enable Slack in any TypingMind chat and the model takes it from there. Its name and description sit in the system prompt, and the moment a request matches, the model loads the full instructions itself — you never invoke it by hand, and it costs no tokens until it is actually used.

The model loads Slack on its own as soon as a request matches it.

Works with any AI model

AI skills are plain Markdown instructions rather than provider-specific code, so Slack is not tied to the model it was written for. Install it once in TypingMind and use it with GPT-5, Claude, Gemini, Grok, DeepSeek, Mistral, Llama, or a local model you run yourself — all on your own API keys.

  • Loaded only when it is needed

    The system prompt carries just the name and description. The instructions are fetched on the first matching request, so an idle skill costs nothing.

  • Switch models mid-chat

    Because the skill is instructions rather than code, changing model does not break it — the next model reads the same SKILL.md.

Skill instructions

This is the SKILL.md content the model loads. Read it before installing — a skill is instructions your model will follow.

Slack

Use Slack as a governed provider boundary without turning Slack—or Runx Cloud—into the owner of the operator's workflow. This skill owns the reusable Slack reading mechanics that an agent should not have to reconstruct: bounded message search, bounded thread hydration, permalink resolution and image reads. Delivering a post or a reply is message-send's job.

The operator-selected provider binding has one narrow role in this flow: retain credential custody and execute a fixed Slack operation. It may be local, self-hosted, third-party, or Runx-hosted. The skill, its procedure, approval point, retry identity, and completion rule remain in Runx OSS. Any queue, team-specific routing rule, or durable action state belongs in a higher-level operator skill.

Runners

search reads one bounded Slack search page. Supply a query object with at least one of author_external_id, mentions_connected_subject, or keywords. The exact fields are:

  • author_external_id: one Slack member id string, such as U123ABC
  • mentions_connected_subject: a boolean selecting messages that mention the member bound to the selected provider account
  • keywords: one keyword-mode search phrase string, not an array and not raw Slack search syntax
  • after and before: ISO-8601 timestamp strings, with after earlier than before
  • channel_types: a non-empty array drawn from public_channel, private_channel, mpim, and im
  • limit: an integer from 1 through 20
  • cursor: the opaque next_cursor from the preceding page

The provider rejects embedded Slack modifiers in keywords, enforces the maximum page, and returns normalized locators and bounded previews rather than a raw Slack response. Continue only from next_cursor; one page is never proof that a workspace scan is complete.

read_thread reads one bounded thread page from an exact slack://workspace/channel/timestamp locator. The limit is capped at 15 because Slack applies a low limit and, for some non-Marketplace installations, a very low request cadence to conversations.replies. Hydrate only threads that are actually needed, preserve next_cursor, and do not fan out speculative reads. Thread messages include bounded attachment metadata and opaque slack-file://workspace/file locators; private Slack download URLs never cross the provider boundary.

read_link accepts one exact Slack archive permalink and reads the same bounded thread page. Use it when an operator pastes a normal https://workspace.slack.com/archives/channel/message link and therefore does not have the workspace id required by a canonical locator. The provider binding validates the permalink, binds it to the authenticated workspace identity, and returns the canonical slack://workspace/channel/timestamp locator. A reply permalink's thread_ts identifies the root; Slack's redundant cid must match the path channel. This runner requires only thread.read, never a search grant.

read_file reads one image attachment from an exact file locator returned by read_thread. It requires file.read, verifies the file through Slack, restricts downloads to Slack's private file origin, accepts images only, and caps decoded bytes at 5 MiB. The result contains the MIME type, content digest, and base64 bytes so the caller can inspect the image without receiving the Slack credential or private URL.

Replies go through message-send: plan_reply binds a proposed reply to the exact canonical thread and its digest, and deliver_reply posts it once after approval and reads it back; reply delivers a reply bound to a source packet. Use message-send send for a proactive top-level channel post. A product-owned skill may add team-specific triage and compose these skills, but must not copy their provider transport, approval, idempotency, readback, or queue logic.

Authority and privacy

Reads resolve a compatible Slack binding for messages.search, thread.read, or file.read and do not ask for human approval. Runx supplies the idempotency key once, and the selected transport verifies the registered operation and its read/mutate class against the bound provider authority.

The skill never receives a Slack token, constructs HTTP, or stores a raw provider envelope. Search and thread outputs contain bounded previews because the operator needs context. Reply receipts contain only stable locators, timestamps, and a content digest; they do not echo the delivered message.

Stop conditions

  • Refuse a missing, ambiguous, revoked, wrong-provider, or insufficient-scope Slack binding. Never fall back to a token, webhook, browser, or provider-specific script.
  • Stop when a search has no structural selector, uses an unsupported modifier, or asks for more than one bounded page per turn.
  • Stop when an attachment is not an image, exceeds 5 MiB, belongs to another workspace, or resolves outside Slack's governed private-file origin.
  • Stop on malformed permalinks, unsupported permalink query fields, cross-workspace locators, or thread/message channel mismatch.
  • Do not infer that a request is resolved from Slack prose. Record resolution, waiting, follow-up, or dismissal explicitly in the owning operator workflow.
  • Do not claim complete scan coverage while a cursor remains.

Worked flow

An operator can search for direct mentions or start from a pasted Slack permalink. They hydrate only the actionable thread, preserve its canonical locator, then pass that normalized observation to a team-specific triage skill. If a reply is needed, message-send plan_reply binds the proposed text to the exact canonical thread. A changed draft fails before approval. An approved unchanged reply is posted once, then read back from the exact returned message locator before Runx seals the effect.

Operator conversation handoff

assistant_scan returns one bounded search page and its continuation cursor; assistant_observe binds each message to its source identity. Hydration checks that identity against the canonical open action before reading the thread. The resulting runx.source_packet.v1 supplies workflow_inputs.thread with the exact source and thread bindings, message evidence and coverage flag, plus source_complete for conversation-review#operate. Missing source messages, truncated content or another page keep coverage incomplete. The consumer holds that work rather than treating a partial thread as resolved.

Bundled files

The model reads these on demand while the skill is loaded. They are exposed as readable files and are never executed.

Frequently asked questions

What does the Slack AI skill do?

Read bounded Slack search, thread, link and file evidence through any compatible Slack binding, ready for an operator's triage and for message-send to reply.

Why use Slack on TypingMind?

Because you install it once and use it with any model. Slack is plain Markdown rather than provider-specific code, so the same skill runs on GPT-5, Claude, Gemini, Grok, or a local model — and you can switch model mid-chat without it breaking. TypingMind runs on your own API keys, so you pay providers directly instead of a per-seat subscription, and your skills and chats stay in your own storage.

How do I install Slack in TypingMind?

Open Plugins → Skills → Install from GitHub in TypingMind and paste https://github.com/runxhq/runx/tree/main/skills/slack. TypingMind reads its SKILL.md and bundles its files and installs it as a skill you can enable per chat.

Which AI models can use Slack?

Any model you connect in TypingMind. AI skills are plain Markdown instructions rather than provider-specific code, so GPT, Claude, Gemini, Grok, and local models can all load this skill when a request matches it.

How many AI models can I use with Slack?

As many as you like. As long as a model supports skills, you can use Slack with it — GPT, Claude, Gemini, Grok, DeepSeek, Mistral, Llama and more — all on TypingMind with your own API keys.

Is the Slack AI skill free?

Yes. It is published on GitHub by runxhq under the Apache-2.0 license. You only pay your own AI provider for the tokens you use.

What are AI skills?

An AI skill is a reusable instruction bundle that teaches an AI model how to do one specific task. It follows the open Agent Skills format: a SKILL.md file with a name and description, plus any scripts, templates or reference files the model may need. The model reads the instructions only when your request matches the skill, so an installed skill costs nothing until it is used.

How are AI skills different from plugins or MCP servers?

A plugin or MCP server gives a model new tools to call — code that runs somewhere and returns a result. An AI skill gives the model knowledge and process instead: how to approach a task, which steps to follow, what good output looks like. Skills are plain Markdown, so they need no server, no API key and no runtime, and they work with any model.

View all

Set up your own AI workspace now

Get notified about new features and future giveaways by subscribing to our newsletter 👇