How iamkk369/project-risk-agent works
Project Risk Agent turns authorized GitHub issue data into deterministic, explainable delivery-risk findings and manager-oriented reports, with optional Gemini executive reasoning. The main workflow fetches and parses issues, scores dependency risk, compares bounded history, enriches decisions, then formats a report; portfolio analysis repeats this per repository. Webhooks can trigger a fresh run, while AgentCore, S3, Secrets Manager, and structured logging support optional runtime integrations. The workflow stages are documented in the README; unsampled implementation details and deployment status are not independently verified here.
Main entry points: main.py CLI Pipeline · github_tool.py Issue Fetch and Parse · webhook.py Webhook Listener · risk_engine.py Risk Engine
Components 18
Risk Analysis
- Issue Fetch and Parse
src/risk_agent/github_tool.py - Risk Engine
src/risk_agent/risk_engine.py - Risk History
src/risk_agent/history.py - Decision Support
src/risk_agent/decision_support.py
Reports and Portfolio
- Markdown Report
src/risk_agent/report.py - Portfolio Analysis
src/risk_agent/portfolio.py
Runtime Integrations
- CLI Pipeline
src/risk_agent/main.py - AgentCore Entrypoint
src/risk_agent/agentcore_app.py - Webhook Listener
src/risk_agent/webhook.py - Event Validation
src/risk_agent/event_handler.py - Encrypted S3 Reports
src/risk_agent/storage.py
Security and Observability
- Credential Resolution
src/risk_agent/security.py - Structured Logging
src/risk_agent/observability.py
Other components
- Engineering Team
- GitHub Issues
- Google Gemini
- Amazon S3
- Secrets Manager
Connections 27
- Engineering Team → CLI Pipeline: starts analysis
README.md - Issue Fetch and Parse → GitHub Issues: fetches issues
README.md - GitHub Issues → Issue Fetch and Parse: returns issue data
README.md - CLI Pipeline → Issue Fetch and Parse: runs pipeline
src/risk_agent/main.py - Issue Fetch and Parse → Risk Engine: parsed issues
README.md - CLI Pipeline → Risk Engine: runs pipeline
src/risk_agent/main.py - Risk Engine → Risk History: current risk snapshot
README.md - CLI Pipeline → Risk History: loads and saves
src/risk_agent/main.py - Risk History → Decision Support: trajectory
README.md - CLI Pipeline → Decision Support: enriches findings
src/risk_agent/main.py - Decision Support → Markdown Report: decision signals
README.md - CLI Pipeline → Markdown Report: generates report
src/risk_agent/main.py - Markdown Report → Engineering Team: presents findings
README.md - CLI Pipeline → Portfolio Analysis: runs portfolio mode
src/risk_agent/main.py - Engineering Team → Portfolio Analysis: requests portfolio
README.md - CLI Pipeline → Google Gemini: requests explanation
README.md - Google Gemini → Engineering Team: explains findings
README.md - Engineering Team → Webhook Listener: starts listener
README.md - Webhook Listener → Event Validation: validates event
src/risk_agent/webhook.py - Webhook Listener → CLI Pipeline: triggers analysis
src/risk_agent/webhook.py - Engineering Team → AgentCore Entrypoint: invokes runtime
README.md - AgentCore Entrypoint → CLI Pipeline: uses workflow
src/risk_agent/agentcore_app.py - CLI Pipeline → Encrypted S3 Reports: persists report
src/risk_agent/main.py - Encrypted S3 Reports → Amazon S3: writes encrypted object
src/risk_agent/storage.py - CLI Pipeline → Credential Resolution: resolves credentials
src/risk_agent/main.py - Credential Resolution → Secrets Manager: reads configured secret
src/risk_agent/security.py - CLI Pipeline → Structured Logging: emits events
src/risk_agent/main.py