An autonomous agent that continuously monitors the Project Management, Program Management, Agile, Engineering Leadership, and Strategy landscape — then delivers structured, actionable intelligence reports tailored for Project Managers, Program Managers, and Tech Leads in the high-tech industry.
| Capability | Details |
|---|---|
| Web intelligence | Polls 30+ curated RSS feeds, GitHub Trending, Arxiv papers, and web sources every 6 hours |
| AI-powered analysis | Uses OpenAI GPT-4o-mini to summarise every article and extract PM-specific insights |
| Trend detection | Clusters articles into trends, calculates momentum scores, and raises alerts for rapidly accelerating themes |
| Self-expansion | Discovers new information sources autonomously by mining article pages for RSS feeds and querying the LLM for recommendations |
| Rich reports | Generates HTML + Markdown reports organised by category, relevance score, trend landscape |
| Notifications | Sends alerts via Email (SMTP) and Slack when high-momentum PM trends are detected |
setup.bat :: creates .venv, installs deps, creates run.bat
run.bat run :: first runWindows note:
run.batsetsPYTHONNOUSERSITE=1to prevent Windows Store Python stubs from shadowing venv packages.
cd pm-intelligence-agent
cp .env.example .env
# Edit .env and set at minimum: OPENAI_API_KEY=sk-...python -m venv .venv
source .venv/bin/activate
pip install -r requirements.txtpython main.py runpython main.py scheduleReports are saved to ./reports/ as HTML files. Open in any browser.
python main.py run # Single run and exit
python main.py schedule # Start recurring scheduler
python main.py schedule --interval 3 # Override interval (hours)
python main.py digest # Run end-of-day digest
python main.py digest --hours 48 # Digest for the last 48 hours
python main.py report # Generate report from existing data
python main.py status # Show run history
python main.py sources # List all registered sourcescp .env.example .env # Fill in OPENAI_API_KEY
docker compose up -d # Start in background
docker compose logs -f # Watch logs
# One-shot run (useful in CI pipelines)
docker compose --profile run-once up pm-agent-run-onceAll settings are via environment variables (.env file):
| Variable | Default | Description |
|---|---|---|
| OPENAI_API_KEY | required | OpenAI API key |
| OPENAI_MODEL | gpt-4o | Model for summarisation and trend analysis |
| OPENAI_MAX_TOKENS | 2000 | Max tokens per summarisation call (600 truncates output) |
| GITHUB_TOKEN | optional | GitHub PAT (raises API rate limit from 60 → 5000 req/hr) |
| SCHEDULE_INTERVAL_HOURS | 6 | How often the agent runs |
| MIN_RELEVANCE_SCORE | 55 | Minimum score (0–100) to include an article in reports |
| NOTIFY_EMAIL | optional | Email address for alert notifications |
| SLACK_BOT_TOKEN | optional | Slack bot token for channel notifications |
| SLACK_CHANNEL | #pm-intelligence | Slack channel to post to |
Each generated report contains:
- Executive Stats – Articles collected, trends detected, alerts count
- 🚨 Alerts – Trends requiring immediate attention (momentum ≥ threshold)
- ⭐ Top 10 Articles – Highest relevance score, with AI summary + PM insights
- 📂 Articles by Category – All articles grouped by: Project Management, Program Management, Agile & Scrum, Engineering Leadership, Strategy & OKRs, AI for PM, PM Tools
- 📈 Trend Landscape – All detected trends with momentum bars and descriptions
| Category | Sources |
|---|---|
| Project Management | PMI Blog, InfoQ PM, Agile Alliance |
| Program Management | SAFe Blog, InfoQ Agile, Medium – Program Management |
| Agile & Scrum | Scrum.org Blog, Mountain Goat Software (Mike Cohn), Martin Fowler, Medium – Agile |
| Engineering Leadership | The Pragmatic Engineer, LeadDev, Will Larson, First Round Review, Lenny's Newsletter |
| Strategy & OKRs | Mind the Product, ProductPlan Blog, Roman Pichler Blog |
| Tools & GitHub | Atlassian Blog, GitHub Blog, GitHub Trending, Dev.to – Management |
| Academic | Arxiv (AI in PM, Agile Research, Effort Estimation, Technical Debt) |
New sources are discovered automatically every run and added to the database.
pm-intelligence-agent/
├── src/
│ ├── agent/ # CoreAgent orchestrator, TrendAnalyzer, SourceDiscoverer, DailyDigestAgent
│ ├── collectors/ # RSS, GitHub, Arxiv, Web scrapers
│ ├── processors/ # RelevanceScorer, Summarizer, ContentProcessor, KeywordExtractor
│ ├── storage/ # SQLAlchemy models + repositories (SQLite/PostgreSQL)
│ ├── reports/ # HTML/Markdown report generator + daily digest generator
│ ├── notifications/ # Email (SMTP) + Slack notifier
│ ├── scheduler/ # APScheduler wrapper
│ └── config/ # Settings (pydantic-settings) + sources.yaml
├── tests/ # Pytest suite (one class per file)
├── reports/ # Generated HTML reports
├── data/ # SQLite database
└── main.py # CLI entry point
CoreAgent.run()
├── 1. Load active sources from DB
├── 2. Collect → RSS + GitHub + Arxiv + Web scrapers
├── 3. Process → Score relevance (fast) → AI summarise (high-score items only)
├── 4. Trends → LLM detects PM/PgM themes across recent articles
├── 5. Discover → Mine pages + ask LLM for new PM source recommendations
├── 6. Report → Generate HTML + Markdown
└── 7. Notify → Email + Slack alerts for high-momentum trends
pytest # Run all tests with coverage
pytest -m "not integration" # Skip tests that need network
pytest tests/test_storage/ # Run a specific moduleOption A – Edit src/config/sources.yaml:
rss_feeds:
- name: "New PM Blog"
url: "https://newpmblog.com/feed"
category: "project_management"
relevance_boost: 10Option B – The agent will discover them automatically based on article links and LLM recommendations.
This agent follows the same architecture as the qa-intelligence-agent in this repository.
Both agents share identical infrastructure patterns (storage, collectors, scheduler, notifications)
but differ in:
| Aspect | QA Intelligence Agent | PM Intelligence Agent |
|---|---|---|
| Focus | GenAI, AI agents, QA & testing | Project/Program Management, Agile, Leadership |
| Sources | Testing blogs, AI labs, DevOps | PMI, Agile Alliance, LeadDev, Pragmatic Engineer |
| Keywords | LLM testing, QA automation, playwright | OKR, Scrum, SAFe, risk management, roadmap |
| Summarizer | QA Manager perspective | Project/Program Manager perspective |
| Categories | genai, agents, qa_testing, devops | project_management, program_management, agile, leadership, strategy |
MIT