How chicken-nuggets-unoffical/the-investigator works
The Investigator is a security-operations copilot with a browser-based workflow and a separate local, automated reporting pipeline. In the app, an analyst can submit logs for correlation, ask questions, browse saved cases, or start an autonomous investigation; the autonomous loop selects evidence and MITRE lookup tools before producing a verdict. Separately, the local triage script combines evidence with the response runbook and saves a report. Standalone scripts provide focused log analysis. The excerpts show these paths, but not the internals of every UI section or deployment configuration.
Main entry points: evidence Evidence logs · app.py Streamlit interface · triage.py Triage runner · agent.py CLI agent loop
Components 21
Web Copilot
- Streamlit interface
app.py - Log correlation
app.py - Analyst chat
app.py - Case browser
app.py
Autonomous Investigation
- Browser agent loop
app.py - CLI agent loop
agent.py - Evidence tools
agent.py - MITRE lookup
agent.py
Local Triage Pipeline
- Triage runner
triage.py - Evidence loader
triage.py - Runbook loader
triage.py - Report writer
triage.py
Focused Log Analysis
- Event timeline
timeline.py - Beaconing analysis
hunt.py - Failed-login audit
audit.py
Case Data
- Evidence logs
evidence - Saved reports
reports - IR runbook
ir_runbook.md
Other components
- Security analyst
- Groq model
- Ollama model
Connections 28
- Security analyst → Streamlit interface: uses
README.md - Streamlit interface → Log correlation: offers
app.py - Streamlit interface → Analyst chat: offers
app.py - Streamlit interface → Case browser: offers
app.py - Streamlit interface → Browser agent loop: offers
app.py - Log correlation → Groq model: sends logs
app.py - Analyst chat → Groq model: sends prompts
app.py - Browser agent loop → Groq model: requests actions
app.py - Browser agent loop → Evidence tools: dispatches tools
app.py - Evidence tools → Evidence logs: reads logs
- Browser agent loop → MITRE lookup: dispatches lookup
app.py - Case browser → Saved reports: reads cases
app.py - Log correlation → Saved reports: downloads report
app.py - Security analyst → CLI agent loop: runs investigation
README.md - CLI agent loop → Groq model: requests actions
agent.py - CLI agent loop → Evidence tools: dispatches tools
agent.py - Evidence tools → Evidence logs: reads logs
agent.py - CLI agent loop → MITRE lookup: dispatches lookup
agent.py - Triage runner → Evidence loader: loads evidence
triage.py - Evidence loader → Evidence logs: reads logs
triage.py - Triage runner → Runbook loader: loads runbook
triage.py - Runbook loader → IR runbook: reads
triage.py - Triage runner → Ollama model: requests report
triage.py - Triage runner → Report writer: saves report
triage.py - Report writer → Saved reports: writes report
triage.py - Event timeline → Evidence logs: reads event logs
- Beaconing analysis → Evidence logs: reads traffic log
- Failed-login audit → Evidence logs: reads access log