daily-review

Generates a prioritized daily status review from agent transcripts, PRs, and weekly plans.

Updated May 12, 2026
One-click install
npx skills add https://github.com/fryanpan/ai-team-lead --skill daily-review-fryanpan
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: daily-review
Source: https://github.com/fryanpan/ai-team-lead/tree/main/.claude/skills/daily-review
Command: npx skills add https://github.com/fryanpan/ai-team-lead --skill daily-review-fryanpan

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? Coordinating multiple AI agents across projects makes it hard to know where things stand each day. This Skill aggregates peer transcripts, hive messages, open PRs, fleet health checks, and the weekly plan into one prioritized, dated review document so you always know what moved, what is blocked, and what needs your decision. ## Core Features & Use Cases - Fleet-wide signal collection: Scans peer transcripts, hive inbox messages, GitHub PRs, Discord, and a fleet healthcheck status file to build the day's picture. - Prioritized review doc: Writes .claude/reviews/YYYY-MM-DD.md with goal sections ordered by weekly-plan priority, plus a decisions-needed section with options, costs, and recommendations. - Automated morning run: Fires on a daily cron before you wake, producing a status summary plus a 2-4 item hit list ordered by goal priority, delivered via a single push notification. - Live-feedback integration: Binds the doc to a review workspace so you can comment and edit inline from your phone, with thread events flowing back as notifications. - Use Case: You ask "where are we?" mid-afternoon. The Skill checks today's transcripts and PRs, updates the review doc, and replies with the workspace URL so you can scan status and leave anchored comments. ## Quick Start Ask the agent to run a daily review and tell you what to focus on today.

Frequently Asked Questions about daily-review

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I get a daily status summary from multiple AI agents?▼

Trigger the daily review by asking "where are we" or "what should I work on". The Skill scans each peer agent's transcripts, hive messages, and open PRs, then writes a prioritized review doc with per-goal status and decisions needed.

How does the automated morning review run work?▼

A team-lead cron fires the Skill each morning around 5:30am local time. It produces a status summary plus a 2-4 item hit list ordered by weekly goal priority and sends one push notification, with no separate messages.

Can I comment on the daily review from my phone?▼

Yes. The review doc is bound to a live-feedback workspace via create_review_doc, and the Skill surfaces a Tailscale workspace URL. You can leave anchored comments and edit inline; thread events arrive back as channel notifications.

What happens when a fleet health check is red?▼

Every RED from /opt/fleet/healthcheck-status.json goes into the review with an owner and an age in consecutive runs. A missing status file is reported as "could not look" and treated as its own RED, never as all-clear.

Why does the review avoid "blocked on" claims without checking transcripts?▼

The Skill greps the peer's session JSONL transcripts for the last processed turn before claiming a blocker, because a tmux pane render cannot show actual session state. This prevents fabricated blocker escalations.

When should I not use this daily review skill?▼

Skip it when a single project with one agent is sufficient and no cross-fleet coordination exists. The Skill is designed for multi-agent fleets with a weekly plan; for simple status checks, reading the board directly is cheaper.