aaaaqwq/daily-briefing-run

Daily Briefing Agent - Manual Run

Runs the daily briefing agent manually (outside of schedule).

When to use

• "run briefing"
• "show morning report"
• "daily briefing"
• Need a report on Sunday/Saturday (outside of schedule)
• Testing after data changes

What it does

The agent collects information from:

• PM tasks (deadline today, in-progress, hot, blocking)
• Email summaries (urgent, reply needed)
• CRM follow-ups (due today + overdue)

Generates:

• AI executive summary (Claude Haiku)
• Priority ranking
• Recommended task to start
• Formatted briefing markdown

Sends to Telegram Saved Messages.

How to execute

Option 1: Dry-run (console, no Telegram)

cd $AGENTS_PATH/daily-briefing
python3 daily-briefing_agent.py --dry-run --verbose

Outputs briefing to stdout, does not send to Telegram.

Use for:

• Viewing report without sending
• Testing after changes
• Debugging errors

Option 2: Full run (with Telegram)

cd $AGENTS_PATH/daily-briefing
python3 daily-briefing_agent.py --now --verbose

Generates and sends to Telegram.

Use for:

• Running on weekends (agent skips Saturday/Sunday)
• Re-running briefing if something was missed
• Testing Telegram integration

Option 3: Tests only

cd $AGENTS_PATH/daily-briefing
python3 test_daily-briefing.py

Runs unit tests without actual execution.

Flags

• --dry-run - generates briefing, outputs to console, does NOT send to Telegram
• --now - bypasses schedule check (allows running on weekends)
• --verbose - detailed logging (shows each step)

What to expect

Successful run:

Step 1: Collecting PM tasks...
Step 2: Collecting email data...
Step 3: Collecting CRM follow-ups...
Step 4: Generating AI summary...
Step 5: Ranking priorities...
Step 6: Formatting briefing...
Step 7: Sending to Telegram...
Telegram notification sent
Step 8: Logging run...
Done! Tasks: 26, Emails: 5, Follow-ups: 3

Error (fallback):

Step 7: Sending to Telegram...
Telegram notification failed: ...
Fallback: saved to /Users/.../data/2026-02-12.md

If Telegram failed -- briefing is saved to a file.

Briefing structure

# Good Morning! Daily Briefing for YYYY-MM-DD

## Executive Summary
- [AI bullet points: critical tasks, emails, follow-ups]

## Email (N new, M actionable)
[URGENT: X | REPLY NEEDED: Y | INFO: Z]

## Tasks Deadline TODAY (N)
- [task] (Project: X) - [description]

## In Progress (N)
- [task] - [notes]

## Hot Tasks (N)
| Score | Task | Description | Note |

## Follow-ups TODAY (N)
[OVERDUE: X if any]
- [Name] @ [Company] via [channel]

## Recommended to Start
**[Task]**
- Why: [reasoning]
- Action: [next step]

Automatic schedule

Agent runs automatically via launchd:

• Mon-Fri at 08:00
• Does NOT run on Saturday/Sunday

Check status:

launchctl list | grep daily-briefing

Logs:

tail -f $AGENTS_PATH/logs/daily-briefing.log
tail -f $AGENTS_PATH/logs/daily-briefing.error.log

Dependencies

Other agents:

• email-monitor (Process #1) - provides email summaries
• Works without email_agent, but Email section will be empty

Tools:

• claude CLI - for AI summary (optional, works without it)
• tg-tools - for Telegram sending

Data:

• $PM_PATH/pm_tasks_master.csv
• $CRM_PATH/contacts/people.csv
• $CRM_PATH/contacts/companies.csv
• $CRM_PATH/activities.csv
• $GOOGLE_TOOLS_PATH/data/email_summaries/YYYY-MM-DD.md

Checking run history

cd $AGENTS_PATH/daily-briefing
python3 -c "
import json
with open('data/agent_log.json') as f:
    log = json.load(f)
    print(json.dumps(log[-5:], indent=2))
"

Shows the last 5 runs with metrics:

• timestamp
• task_count, email_count, followup_count
• recommended_task_id
• status (success/partial/error)
• errors[]

Troubleshooting

"No such file or directory: pm_tasks_master.csv"

• File is missing or path is incorrect
• Agent will continue with empty tasks

"Email summary not available for today"

• Normal if email_agent has not run yet (before 09:00)
• Or email_agent is broken -- check its logs

"Claude CLI error"

• Check: echo "test" | claude -p --model haiku
• Agent will skip AI summary, show structured data

"Telegram notification failed"

• Check Telegram session: ls /Users/.../telegram/sessions/*.session
• Briefing is saved in data/YYYY-MM-DD.md

Output

• Telegram: Saved Messages (if successful)
• Fallback: $AGENTS_PATH/daily-briefing/data/YYYY-MM-DD.md
• Log: $AGENTS_PATH/daily-briefing/data/agent_log.json

Related skills

• email-monitor - run email agent
• git-workflow - if agent code was modified
• pm/task-prioritize - manual task prioritization

Owner

Your Name ([email protected])

Process ID: #14 Agent type: Scheduled reporting (read-only)