agentleFS
egregore-labsegregore-labs/egregore/.claude/skills/slack-connect/SKILL.md

slack-connect

Skill293 starsChanged 11 days agoMITView on GitHub

Set up Slack as a notification channel for your Egregore — creates the workspace app, invites the bot, and binds a channel. Use for /slack-connect, 'connect slack', or 'set up slack notifications'.

  • Reads credentials

The file

We split this skill into the rules it gives an agent

11 rules. About 0 are common, 4 are distinctive, and 7 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈4

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 7

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  • could be automatic: a setting could enforce it
  1. Distinctive:

    Works in both modes: Check which one first:

    line 34
  2. Very specific:

    Use the printed value as {mode}.

    line 39
  3. Distinctive:

    Only Step 5 differs between the modes.

    line 46
  4. Distinctive:

    Do not store it anywhere yet.

    line 76
  5. Very specific:

    Ask for the channel ID: channel name → About tab → Channel ID at the bottom (or the /archives/C… segment of a channel link).

    line 86
  6. Very specific:

    Write the token they pasted in Step 2 to .env (gitignored — never commit) before testing it: append one line, SLACK_BOT_TOKEN= followed by …

    line 92
  7. Very specific:

    Only read the response with jq after the helper succeeds.

    line 131
  8. Very specific:

    Write the file tmp/slack-connect-bind.json with {"channel_id":"<CHANNEL_ID>","bot_token":"<TOKEN>","channel_name":"<#channel-name>"}, …

    line 135
  9. Very specific:

    Follow .claude/context/notification-consent.md.

    line 152
  10. Very specific:

    Use the printed values as {plan_id} and {digest}, respectively, for the notification consent flow.

    line 166
  11. Distinctive:

    Only dispatch after Send is selected for that preview.

    line 175

≈ “Common” is an estimate: at least 1 in 5 of the 372,429 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Set up Slack as a notification channel for your Egregore.

Arguments: $ARGUMENTS (none expected)

When to invoke

Trigger phrases:

  • "set up slack", "connect slack", "slack-connect"
  • "add slack notifications", "slack channel", "notify us on slack"

Disambiguation:

  • "set up telegram" → /telegram-connect
  • "set up teams" → /teams-connect
  • "send a message to slack" → NOT this — use the separate notification consent protocol in .claude/context/notification-consent.md

What this does

Connects one Slack channel to Egregore notifications using an app the org creates in its own workspace. After this, everything that plans a group notification may also resolve Slack as a receiving channel; all resolved channels are shown before separate dispatch approval.

The privacy contract, stated plainly to the user: the install screen will say only "Send messages as @Egregore." The app cannot read messages — the manifest requests one scope, chat:write. The workspace owns the app and the token; revoking either cuts Egregore off instantly.

Works in both modes. Check which one first:

bash bin/config-get.sh mode

Use the printed value as {mode}; quote it in single quotes when you use it in a command, and write any single quote inside the value as '\''.

  • local — the token stays in the gitignored .env on this machine and messages go straight from here to Slack. No Egregore server is in the path.
  • connected — the token is stored on the org's row so the hosted service can deliver group notifications from any member's session.

Only Step 5 differs between the modes.

Step 0: Prerequisites

  1. A Slack workspace where this person can install apps. Some workspaces require admin approval for new apps — if theirs does, the install in Step 2 turns into a request the admin must approve first.
  2. Nothing to install locally. No CLI, no cloud account.

Step 1: Create the app from the manifest (human, ~1 min)

Show the manifest:

cat slack-app/manifest.yml

(If this instance doesn't carry slack-app/, fetch it from the framework repo.)

Guide them: https://api.slack.com/apps → Create New App → From a manifest → pick their workspace → paste the manifest → Create.

Step 2: Install and copy the bot token (human, 2 clicks + 1 paste)

Settings → Install App → Install to Workspace → Allow.

Then copy the Bot User OAuth Token from the same page. It starts with xoxb-. Validate the prefix when they paste it:

  • xoxb- — correct.
  • xoxp- (user token) or xapp- (app-level token) — wrong item on the page; point them back to Install App → Bot User OAuth Token.

Do not store it anywhere yet.

Step 3: Invite the bot to the channel (human, 1 line)

In the target channel: /invite @Egregore (or channel name → Integrations → Add apps). With only chat:write, the bot can post solely in channels it is a member of — skipping this yields not_in_channel at send time.

Step 4: Channel ID (human pastes, agent validates)

Ask for the channel ID: channel name → About tab → Channel ID at the bottom (or the /archives/C… segment of a channel link).

Validate: must match ^[CG][A-Z0-9]{6,}$. A D… id is a DM — not supported; ask for a channel.

Write the token they pasted in Step 2 to .env (gitignored — never commit) before testing it: append one line, SLACK_BOT_TOKEN= followed by the pasted value, by editing the file directly, never through a shell command, so the value does not enter the command log. This is the one place the pasted value is written; the smoke test reads it from there instead of taking it as a typed argument.

Smoke-test the stored token (no message is sent):

bash bin/notify.sh slack-auth-test

"ok":true proves the token and names the workspace — confirm it is the one they expect. invalid_auth here means a mispasted or revoked token; have them paste it again, rewrite the SLACK_BOT_TOKEN line in .env, and re-run. Fix now, not at send time.

Step 5: Persist

Local mode — channel to .egregore-state.json (NOT egregore.json — the state file is untracked and shared across worktrees):

mkdir -p tmp
jq '.slack_channel_id = "<CHANNEL_ID>" | .slack_channel_name = "<#channel-name>"' \
  .egregore-state.json > tmp/slack-connect-state.json

After jq succeeds:

mv tmp/slack-connect-state.json .egregore-state.json

Connected mode — store on the org so any member's session can deliver:

mkdir -p tmp

Only read the response with jq after the helper succeeds. If it fails, show its error and stop; the helper leaves no response file after a failure.

Write the file tmp/slack-connect-bind.json with {"channel_id":"<CHANNEL_ID>","bot_token":"<TOKEN>","channel_name":"<#channel-name>"}, substituting the channel details and the token the user pasted. The helper reads EGREGORE_API_KEY internally.

bash bin/api-call.sh POST /api/org/slack --auth egregore --json-file tmp/slack-connect-bind.json --out tmp/slack-connect-bind-response.json
jq '.' tmp/slack-connect-bind-response.json

Expect {"status":"connected", …}.

Step 6: Verify with separate exact notification consent

Follow .claude/context/notification-consent.md. Connecting Slack is not consent to send a test message. Prepare after configuration is final:

mkdir -p tmp
bash bin/notify.sh plan group "Slack channel connected — this message reached you through Egregore." > tmp/slack-connect-plan.json
jq -r '.plan_id, .digest' tmp/slack-connect-plan.json

Use the printed values as {plan_id} and {digest}, respectively, for the notification consent flow. Read the exact preview fields:

jq -r '.org, .recipient, .channels, .deliveries, .message' tmp/slack-connect-plan.json

Show every resolved receiving channel—potentially Telegram, Teams, and Slack—and the exact message in a dedicated Send / Edit / Cancel checkpoint. Only dispatch after Send is selected for that preview.

Troubleshooting quick table

SymptomCause → fix
invalid_authWrong, revoked, or other-workspace token → Step 2, re-copy from Install App
not_in_channelBot not invited to the channel → Step 3
channel_not_foundWrong channel ID, or a private channel the bot isn't in → Step 4 / invite it
missing_scopeApp not created from the manifest → recreate from Step 1, reinstall
account_inactiveApp was uninstalled from the workspace → reinstall
ratelimitedTransient — retried automatically
Install becomes an approval requestWorkspace requires admin approval → the admin approves, then continue
Local plan omits slack from channelsHalf-configured: SLACK_BOT_TOKEN missing from .env or slack_channel_id missing from .egregore-state.json

After

Emit telemetry (fire-and-forget):

bash bin/telemetry.sh emit "slack-connect" '{"command":"slack-connect"}' 2>/dev/null &

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.