Weekly Report Skill

The Weekly Report skill packages the FreightTiger TMS leadership report workflow so another Codex agent or FreightTiger operator can generate the report consistently from Pylon data.

Downloads:

• FreightTiger weekly report skill
• Standalone report pipeline

Why It Was Created

FreightTiger leadership needs a weekly view of TMS customer health, issue load, SLA risk, module quality, and team actions. The skill makes the reporting workflow repeatable instead of relying on a person remembering the scripts, filters, caveats, and publish steps.

Business goals:

• Show which accounts need attention this week.
• Highlight SLA breaches and aged open bugs.
• Identify module-level product or engineering pressure.
• Create persona-specific actions for Account Managers, Engineering, and Product.
• Make the pipeline installable in FreightTiger's own cloud/system.

User Story

Every week, an operator should be able to run one documented workflow that pulls Pylon data, computes metrics, generates summaries, renders a leadership report, and optionally publishes it. Once FreightTiger hosts this pipeline, implementation-owned workers and scheduled jobs can be shut down.

How It Works

The skill bundles a runnable pipeline:

Stage	Script	Purpose
Pull	01_pull_pylon.py	Pull TMS accounts and 28 days of Pylon issues.
Metrics	02_compute_metrics.py	Compute SLA, health, breach, module, account, tier, and red-flag metrics.
Summaries	03_generate_summaries.py	Generate LLM concerns and persona actions using local Ollama.
Render	04_render_report.py	Render HTML and Markdown reports.
Publish	05_publish.sh	Encrypt and publish the report archive to Cloudflare Pages.

See Analytics Logic for metric details.

Step-By-Step Use

1. Install The Skill

mkdir -p ~/.codex/skills
tar -xzf freighttiger-weekly-analytics-skill.tar.gz -C ~/.codex/skills

Invoke it:

Use $freighttiger-weekly-analytics to generate this week's FreightTiger TMS report from Pylon.

2. Install The Runtime Pipeline

tar -xzf freighttiger-weekly-analytics-pipeline.tar.gz
cd report-pipeline
python3 -m venv .venv
.venv/bin/pip install -r requirements.txt
cp .env.example .env

Set:

PYLON_API_KEY=...

3. Start The Summary Model

ollama serve
ollama pull qwen3:8b

4. Run The Report

make all

Expected outputs:

data/weekly_snapshots/{date}/...
reports/{date}/report.html
reports/{date}/report.md

5. Publish If Needed

make cf-login
make publish REPORT_PASSWORD='...' PAGES_PROJECT='ft-tms-weekly'

Hosting In FreightTiger Cloud

Recommended weekly cron:

cd /opt/ft-weekly-analytics
make all
make publish REPORT_PASSWORD="$REPORT_PASSWORD" PAGES_PROJECT="ft-tms-weekly"

Store these outside source control:

• PYLON_API_KEY
• REPORT_PASSWORD
• Cloudflare credentials

Handoff Note

Once FreightTiger deploys and schedules this pipeline in their own environment, implementation-owned scheduled jobs and publishing workers should be shut down.