Menu
Agent → CEO report. Send a status report or question to the user (CEO) on behalf of an agent. Notifies the dashboard inbox and, when a Slack `ceoChannel` is configured, pushes a Slack message.
Install with Codex or Claude Copy this prompt, paste it into Codex, Claude, or another assistant, and let it review the skill page and install it for you.
Based on SOC occupation classification
| name | report |
| description | Agent → CEO report. Send a status report or question to the user (CEO) on behalf of an agent. Notifies the dashboard inbox and, when a Slack `ceoChannel` is configured, pushes a Slack message. |
| trigger | report to ceo, agent report, ask ceo, status report, escalate to user, aweek report |
Send a CEO report on behalf of an agent. Every report:
.aweek/agents/<slug>/notifications.json, which surfaces immediately on the dashboard's inbox..aweek/channels/slack/config.json (or env) provides BOTH a botToken AND a ceoChannel, posts a formatted Slack message to that channel/DM via chat.postMessage.The skill is the canonical surface agents call autonomously during a heartbeat session (aweek exec report send …) AND the user-facing slash command for manually composing or testing a report. Use aweek:notify directly only when you need the generic notification surface without the report/question framing.
During a heartbeat session an agent decides "I should tell the CEO about this" or "I'm blocked and need a decision" — it shells out to:
echo '{
"senderSlug": "<AGENT_SLUG>",
"kind": "report",
"title": "<SHORT_TITLE>",
"body": "<DETAILED_BODY>",
"severity": "info"
}' | aweek exec report send --input-json -
Replace kind with "question" when the agent needs a decision (rather than a status update). Replace severity with "warning" (action recommended) or "error" (blocked / urgent).
You do not invoke this on the agent's behalf — agents call it themselves through Bash during their heartbeat tick. The slash command below is for the USER to manually compose a report (testing, manual escalation, or surfacing context the agent can't articulate yet).
You MUST follow this exact workflow when invoked. Use aweek exec report send — never write notification JSON files directly.
List available agents so the user can choose which one the report should be attributed to:
echo '{"dataDir":".aweek/agents"}' \
| aweek exec agent-helpers listAllAgents --input-json -
Project each entry to { id, name, role } for display. Treat zero entries as the "create an agent first" case — direct the user to aweek hire and stop.
Ask the user via AskUserQuestion: "Which agent is sending the report?"
Ask the user via AskUserQuestion: "What kind of message is this?"
report — status update; no decision needed.question — agent is blocked / needs a decision from the CEO.AskUserQuestion: "Title for the report (≤ 200 chars):"AskUserQuestion: "Body (≤ 5000 chars). Be specific — this is what the CEO reads:"AskUserQuestion: "Severity? (info / warning / error — default info)"AskUserQuestion: "Link this report to a weekly-task ID? (press Enter to skip)"Display a summary in the conversation (do not rely on Bash output rendering — re-emit as direct assistant text):
--- Report Summary ---
From: <AGENT_NAME> (<AGENT_ID>)
Kind: <KIND>
Severity: <SEVERITY>
Title: <TITLE>
Body: <BODY>
Task: <SOURCE_TASK_ID or "none">
Ask via AskUserQuestion: "Send this report?"
If confirmed, execute:
echo '{
"senderSlug": "<AGENT_ID>",
"kind": "<KIND>",
"title": "<TITLE>",
"body": "<BODY>",
"severity": "<SEVERITY>",
"sourceTaskId": "<SOURCE_TASK_ID_OR_OMIT>"
}' | aweek exec report send --input-json -
Omit severity and sourceTaskId from the JSON when the user skipped them. JSON-escape every interpolation.
The response is the { notification, deliveredToSlack, persisted } envelope. Re-emit the formatted summary as direct assistant output:
echo "$RESULT" | aweek exec report formatReportResult --input-json - --format text
Highlight:
notification.id — for traceability in .aweek/agents/<slug>/notifications.json and on the dashboard inbox.deliveredToSlack — reflects the actual delivery outcome: true iff Slack returned ok: true for the chat.postMessage. false means one of (a) no ceoChannel was configured so the Slack push was deliberately skipped (only the dashboard inbox got the write — instruct the user to set one if they want Slack push), (b) Slack returned ok: false (the error code — channel_not_found, not_in_channel, invalid_auth, etc. — appears in STDERR), or (c) the network/fetch itself failed. The skill awaits the Slack POST before returning, so by the time you read this field the message has either landed in Slack or visibly failed there is no longer an "in-flight maybe" window.If the user declines, say "Report cancelled. No notification persisted, no Slack message sent." and stop.
The Slack push is opt-in per project. To enable it:
aweek slack-init (creates the Slack app and persists botToken / appToken into .aweek/channels/slack/config.json).ceoChannel to the same file — a Slack channel ID (C…/G…), user ID (U…), or DM ID (D…). The bot must already be a member of channels (C…/G…); DMs (D…) and user IDs (U…) work without membership.The file ends up like:
{
"botToken": "xoxb-…",
"appToken": "xapp-…",
"signingSecret": "…",
"ceoChannel": "D01ABCXYZ"
}
Env-vars take precedence per the loader contract: SLACK_CEO_CHANNEL overrides the file's ceoChannel.
When ceoChannel is missing, the report still persists to the dashboard inbox — only the Slack push is skipped. The agent's aweek exec report send returns deliveredToSlack: false so the agent knows.
senderSlug must match an existing agent in .aweek/agents/.kind must be report or question exactly.title non-empty, ≤ 200 chars.body non-empty, ≤ 5000 chars.severity, when present, must be info / warning / error.channel_not_found) does NOT roll back the notification; the failure is logged to stderr by the dispatcher's per-call channel-error sink.Show the project's `.aweek/config.json` knobs (plus the hardcoded scheduler/lock constants the dashboard surfaces) and edit the config-backed fields with a confirmation gate
Edit an agent's free-form planning markdown and adjust weekly tasks from a single entry point
Manage Google-Calendar-style recurring tasks per agent — daily/weekly/biweekly/monthly rules that expand into occurrences on the calendar and execute through the heartbeat loop
Bootstrap an aweek-branded Slack app and persist its credentials so the embedded SlackAdapter inside `aweek serve` can chat with the project's Claude through Slack
Set up an aweek project — create the data directory and optionally install the scheduled heartbeat (launchd on macOS, crontab elsewhere)
Remove the aweek heartbeat and/or the .aweek/ data directory from a project