An LLM agent-based reporting system that provides insights into ongoing Greymass and Wharfkit development activities. The system automatically collects data from GitHub and blogs, then generates reports - no need to understand GitHub or browse through hundreds of projects.

View these reports online: reports.greymass.io

The website provides a more readable experience with a timeline view, navigation, and better formatting.

A note from a human: We're experimenting with this as a way for everyone to understand the work we're doing. It may be inaccurate though. This information is reviewed before being added to this repo - but we may not catch everything. Let us know if anything looks incorrect.

• February 2026 - Latest report - Session keys, contract browser, prototypes (technical | research)
• January 2026 - Roborovski open-sourced, sentiment expansion (technical | research)
• December 2025 - Governance voting, session keys (technical | research)
• November 2025 - Sentiment polling, governance tools (technical | research)
• October 2025 - Passkey integration, backup prototypes (technical | research)
• September 2025 - Testnet refinements, Unicove integration (technical | research)
• August 2025 - Backup prototypes, token contracts (technical | research)
• July 2025 - Web Authenticator design, Spring testnet (technical | research)
• June 2025 - Transaction signing, Sextant v2 (technical | research)
• May 2025 - Vaulta rebrand, light theme (technical | research)
• April 2025 - Token Swap, Vaulta rebrand (technical | research)
• March 2025 - Translations, token refactoring (technical | research)
• February 2025 - Multisig support, Web Authenticator deployment (technical | research)
• January 2025 - Unicove contracts, account tracking (technical | research)
• December 2024 - Unicove 2.0, EOS Wallet launch (technical | research)
• November 2024 - Unicove launch preparation (technical | research)
• October 2024 - Unicove refinements, MetaMask audit (technical | research)
• September 2024 - Resource management features (technical | research)
• August 2024 - Core wallet features (technical | research)
• July 2024 - Unicove 2.0 development start (technical | research)
• June 2024 - Anchor Mobile improvements (technical | research)

Generates comprehensive activity reports using a multi-agent LLM architecture:

• Executive summaries for stakeholders (plain language, non-technical)
• Technical reports for developers (detailed with GitHub links)
• Research data collected automatically from GitHub, blogs, and publications

The goal is to make development activity transparent and understandable without requiring readers to navigate GitHub repositories or understand technical jargon.

The system uses a Director-Agent pattern with specialized LLM sub-agents:

1. Director (agent.md) - Orchestrates the workflow
2. Researcher (agents/researcher.md) - Collects data from GitHub/RSS feeds
3. Summary Writer (agents/summary.md) - Creates executive summaries
4. Technical Writer (agents/technical.md) - Creates technical reports
5. Report Fixer (agents/report-fixer.md) - Applies targeted fixes to existing reports

• agent.md - Director agent (orchestrates report generation)
• agents/ - Specialized sub-agents
  • researcher.md - Data collection from GitHub/RSS
  • summary.md - Executive summary writer
  • technical.md - Technical report writer
• repo-names.md - Maps repository names to friendly display names
• reports/ - Generated reports directory
  • Each report is a folder (e.g., 2025-10/) containing:
    • README.md - Executive summary (non-technical)
    • technical.md - Technical details (for developers)
    • Optional: Additional files from other agents

Simple approach: Ask the Director agent to generate a report

"Generate a report for October 2025" 

The Director will:

1. Invoke the Researcher to collect data
2. Present findings for your approval
3. Invoke Summary Writer to create README.md
4. Invoke Technical Writer to create technical.md
5. Update this file with the new report link

Manual approach: Invoke agents individually

1. Use Researcher agent to collect data
2. Use Summary Writer with research data
3. Use Technical Writer with research data