Shannon provides multiple ways to monitor penetration test execution in real-time: workflow logs, Temporal Web UI, and direct queries.Documentation Index
Fetch the complete documentation index at: https://mintlify.com/KeygraphHQ/shannon/llms.txt
Use this file to discover all available pages before exploring further.
Quick Monitoring Commands
Workflow Logs
The primary way to monitor Shannon execution is through workflow logs.Viewing Logs
Log Format
Workflow logs are human-readable and show:- Timestamps - When each event occurred
- Phase transitions - Moving between Pre-Recon, Recon, Analysis, etc.
- Agent status - Starting, running, completed, failed
- Tool usage - Browser actions, API calls, file operations
- Errors - Failures, retries, timeouts
Log File Location
Workflow logs are saved to:OUTPUT=), logs are saved to:
Auto-Discovery
The./shannon logs command automatically finds log files:
- Checks default location:
./audit-logs/<id>/workflow.log - Handles resume workflow IDs:
workspace_resume_123 - Searches custom output directories
- Falls back to
findfor non-standard locations
- Named Workspace
- Auto-Named Workspace
- Custom Output
Temporal Web UI
Shannon uses Temporal for workflow orchestration. The Temporal Web UI provides detailed execution visibility.Accessing the UI
Open in your browser:For WSL2 UsersIf running Shannon in WSL2, find your WSL IP address:Then navigate to
http://<wsl-ip>:8233 from your Windows browser.UI Features
Workflow List
View all running and completed workflows
Execution History
Timeline of all events and activities
Activity Details
Input/output for each agent execution
Error Inspection
Stack traces and failure details
Finding Your Workflow
Open Temporal UI
Navigate to http://localhost:8233
Key Views
- History
- Pending Activities
- Stack Trace
- Query
Timeline of all events:
- Workflow started
- Activity scheduled
- Activity started
- Activity completed
- Activity failed
- Workflow completed
Workflow Queries
Shannon supports querying workflow state programmatically, but this is only available through the Temporal Web UI or custom code.No CLI Query CommandShannon does not currently provide a
./shannon query command. To check workflow progress:- Use the Temporal Web UI at
http://localhost:8233 - Use
./shannon logs ID=<workflow-id>to view detailed logs - Use
./shannon workspacesto list all workspaces and their status
PROGRESS_QUERY that returns:
src/temporal/shared.ts for the query definition and src/temporal/workflows.ts for query handlers.
Workspace Listing
View all workspaces with status, duration, and cost:Status Indicators
Workflow is currently executing
All agents finished successfully
Workflow encountered an error or was interrupted
Workspace Metadata
Each workspace directory containssession.json with execution metadata:
Agent Logs
Individual agent execution logs are saved to:- Detailed tool usage
- LLM prompts and responses (excerpts)
- Error details
- Retry attempts
Viewing Agent Logs
Monitoring Best Practices
Keep Logs Running
Open
./shannon logs in a separate terminal window to monitor progress in real-time:Check Temporal UI Periodically
Use the Temporal Web UI to:
- Verify heartbeats are active (no stuck activities)
- Check for retry patterns (may indicate issues)
- Inspect failed activities for detailed errors
Troubleshooting
Logs not found for workflow ID
Logs not found for workflow ID
Error:Causes:
- Workflow hasn’t started yet (check Temporal UI)
- Incorrect workspace ID (use
./shannon workspacesto list) - Custom output directory not searched (specify full path)
./shannon workspaces or check Temporal UI.Temporal UI not accessible
Temporal UI not accessible
Issue: Can’t access http://localhost:8233Causes:
- Containers not running
- Port conflict on 8233
- WSL2 networking (use WSL IP instead of localhost)
Agent stuck (no log updates)
Agent stuck (no log updates)
Issue: Logs stop updating, agent appears frozenCheck in Temporal UI:
- Pending activities with long duration (>30min)
- Heartbeat status (should be periodic)
- Failed heartbeats (indicates stuck activity)
- Wait for timeout/retry (automatic)
- Manual stop and resume if timeout is too long
High API costs
High API costs
Issue: Costs increasing rapidlyCheck:
session.jsonfortotal_cost_usd- Agent logs for excessive retries
- Temporal UI for failed activities
- Reduce
max_concurrent_pipelinesin config - Use smaller models (if overridden)
- Stop and review before continuing
Monitoring Checklist
Before starting a production audit:Open Temporal UI
Ensure http://localhost:8233 is accessible
Related Resources
Resume Workflow
Resume interrupted runs from checkpoints
Configuration
Adjust retry and concurrency settings