The fourth repository in the SportMind suite. Monitors, classifies, and delivers structured intelligence briefings. You decide what enters the library.
A standalone, forkable briefing system that watches defined public sources on a schedule, classifies findings against the SportMind three-tier intake framework, and delivers a structured intelligence briefing with full reasoning.
It does not push to GitHub autonomously. It does not modify the library. It researches, classifies, and presents. You decide what gets added.
Every briefing item is mapped to a specific SportMind file and modifier. Signals with no clear library mapping are filtered before delivery. What reaches you is actionable, not noise.
| Repository | Role |
|---|---|
| SportMind/SportMind | Core library — the intelligence primitive set |
| SportMind/mcp-server | MCP server — exposes library to AI agent runtimes |
| SportMind/agent-examples | Reference agent implementations |
| SportMind/calibration-registry | Public calibration record archive |
| SportMind/intelligence-agent | This repository — monitoring, classification, briefing |
Monitor sources → Filter noise → Classify (Tier 1/2/3) → Map to SportMind files → Reason → Deliver briefing
- Monitor — watches public RSS feeds, free-tier APIs, and public channels across five source categories
- Filter — discards signals with no clear SportMind file mapping
- Classify — assigns Tier 1, 2, or 3 using the intake framework
- Map — identifies which specific SportMind files and modifiers are affected
- Reason — generates a structured reasoning chain for each item
- Deliver — opens a GitHub Issue in
SportMind/SportMindtitledSportMind Intelligence Briefing — [date]
See agent/agent-pattern.md for the full reasoning chain.
- Fork this repository
- Add your GitHub token as a repository secret:
SPORTMIND_GH_TOKEN - Edit
sources/default-sources.mdto enable/disable source categories - The workflow in
examples/github-actions-workflow.ymlruns automatically on schedule
The default schedule is weekly on Monday at 07:00 UTC.
git clone https://github.com/SportMind/intelligence-agent
cd intelligence-agent
python agent/run.py --output terminalThe briefing format and classification framework are structured markdown. Load them into any agent runtime alongside the SportMind core library. See briefing/briefing-format.md.
| Method | Best for | Setup |
|---|---|---|
| GitHub Issue | Auditable public record (default) | Add SPORTMIND_GH_TOKEN secret |
| Telegram | Real-time mobile alerts | Add TELEGRAM_BOT_TOKEN + TELEGRAM_CHAT_ID |
| Email digest | Weekly summaries | Add SMTP credentials |
Full setup: briefing/delivery-options.md
- Regulatory — SEC/CFTC filings, MiCA updates, CoinDesk regulatory coverage
- Chiliz ecosystem — Socios announcements, CHZ on-chain data, fan token governance
- Sports calendar — competition schedules, transfer windows, fixture announcements
- Macro crypto — Bitcoin cycle signals, exchange listings, market structure events
- Fan token specific — new launches, supply events, burn/mint confirmations
Full source list: sources/default-sources.md
| Tier | Name | Action |
|---|---|---|
| Tier 1 | Primary signal | Immediate library update candidate |
| Tier 2 | Corroborating signal | Strengthens existing intelligence |
| Tier 3 | Background context | Monitoring only, no immediate action |
Full framework: classification/intake-framework.md
See examples/sample-briefing.md for a complete worked example.
See CONTRIBUTING.md.
- Works alongside all four other SportMind suite repositories
- No external dependencies beyond Python standard library +
requests - Compatible with any LLM
- Classification framework maps directly to
core/external-intelligence-intake.mdinSportMind/SportMind
MIT — see LICENSE
SportMind/intelligence-agent v1.0.0