Agent Reports
Generate comprehensive activity reports for your AI agents, complete with session replay, risk assessment, and exportable output in text, markdown, or JSON format.
Usage
suv agent reportThis generates an activity report covering agent commands from the specified time range. By default, it covers today's activity.
Flags
| Flag | Description |
|---|---|
--after <TIME> | Start of the report period (default: today). Accepts natural language like "3 days ago" or dates like "2025-01-15". |
--before <TIME> | End of the report period. Accepts the same time formats as --after. |
--executor <NAME> | Generate report for a specific agent only (e.g., claude-code, cursor) |
--format <FORMAT> | Output format: text (default), markdown, or json |
--here | Only include commands from the current directory |
What the Report Includes
Session Replay
A chronological walkthrough of every agent session in the report period. Each session shows the commands in order, with timestamps, working directories, and exit codes — so you can replay exactly what happened.
Risk Levels
Every command in the report is annotated with its risk level (Critical, High, Medium, Low, or Safe). This makes it easy to scan for potentially dangerous operations without reading every command.
Risk Summary
An aggregate breakdown showing how many commands fell into each risk category across the entire report period. This gives you a quick sense of whether agent activity was routine or involved significant risk.
Command Listing
A complete listing of all agent commands with metadata: executor name, timestamp, directory, exit code, duration, and risk level. In JSON format, this is a structured array suitable for further processing.
Output Formats
Text (default)
Human-readable plain text output, printed directly to stdout. Suitable for quick review in the terminal.
suv agent reportMarkdown
Formatted markdown with headers, tables, and code blocks. Suitable for pasting into GitHub issues, pull request descriptions, Notion pages, or any markdown-aware tool.
suv agent report --format markdownJSON
Structured JSON output with all data fields. Suitable for programmatic processing, piping to jq, or feeding into other analysis tools.
suv agent report --format jsonExamples
Today's activity as markdown
suv agent report --format markdownGenerates a markdown report of all agent activity today. Pipe to a file with > report.md to save it.
Claude Code activity over the last 3 days as JSON
suv agent report --after "3 days ago" --executor claude-code --format jsonGenerates a JSON report covering only Claude Code commands from the past 3 days. Useful for automated analysis or dashboards.
Report for a specific date range
suv agent report --after "2025-01-01" --before "2025-01-31"Generates a text report for all agent activity in January 2025.
Current project activity
suv agent report --here --format markdownGenerates a markdown report scoped to the current working directory, useful for project-specific agent audits.