Skip to main content

Activity Feed

The Enterprise Agent Activity Feed provides comprehensive monitoring of all AI agent actions with CVSS scoring, MITRE ATT&CK mappings, and approval workflow tracking.

Overview

View all agent actions with detailed security assessments, compliance mappings, and approval status in a unified interface.

Source: owkai-pilot-frontend/src/components/AgentActivityFeed.jsx

Compliance: SOC 2 AU-2, NIST AU-2, PCI-DSS 10.2

Activity Card

Activity Information

Each activity displays:

┌─────────────────────────────────────────────────────────────┐
│ financial-advisor-001 CVSS: 6.5 (MEDIUM) PENDING │
├─────────────────────────────────────────────────────────────┤
│ Action Type: database_write │
│ Tool: postgres_connector │
│ Description: Update customer profile with new preferences │
├─────────────────────────────────────────────────────────────┤
│ 🕐 Dec 7, 2025, 10:30 AM 👤 User ID: usr_123 │
├─────────────────────────────────────────────────────────────┤
│ [▶ Expand Details] [⚠ Mark False Positive] [🔁 Replay] │
└─────────────────────────────────────────────────────────────┘

Basic Fields

FieldDescription
agent_idAgent identifier
action_typeType of action performed
tool_nameTool/connector used
descriptionAction description
timestampWhen action occurred
user_idAssociated user (if any)

Filters

Available Filters

FilterOptions
Risk LevelAll, High, Medium, Low
StatusAll, Pending, Approved, Denied, In Review
SearchAgent, tool, or description text

Filter Usage

// API query with filters
GET /api/v1/actions?risk=high&status=pending

Expanded Details

Security Assessment Card

FieldDescriptionStandard
cvss_scoreCVSS 3.1 score (0.0-10.0)CVSS v3.1
cvss_severitySeverity levelCVSS v3.1
cvss_vectorAttack vector stringCVSS v3.1
risk_scoreInternal risk (0-100)OW-kai

MITRE ATT&CK Mapping

FieldDescriptionExample
mitre_tacticAttack tacticExfiltration
mitre_techniqueTechnique IDT1041

NIST Control Reference

FieldDescriptionExample
nist_controlControl IDSI-4
nist_descriptionControl nameInformation System Monitoring

Approval Workflow Card

FieldDescription
statusCurrent approval status
current_approval_levelCurrent level in workflow
required_approval_levelTotal levels required
approved_byWho approved (if applicable)
reviewed_byWho reviewed
reviewed_atReview timestamp
pending_approversAwaiting approval from
sla_deadlineResponse deadline

Target Details Card

FieldDescription
target_systemTarget system name
target_resourceTarget resource path

AI Summary

When available, displays AI-generated action summary for quick understanding.

CVSS Scoring

Severity Levels

Score RangeSeverityColor
9.0 - 10.0CriticalRed
7.0 - 8.9HighOrange
4.0 - 6.9MediumYellow
0.1 - 3.9LowGreen
0.0NoneGray

CVSS Badge Display

CVSS: 6.5 (MEDIUM)

Actions without CVSS assessment display "No CVSS" badge.

Status Badges

StatusColorDescription
PendingYellowAwaiting review
ApprovedGreenAction approved
DeniedRedAction denied
In ReviewBlueUnder investigation

Actions

Mark as False Positive

  1. Click Mark as False Positive
  2. Action is flagged for model improvement
  3. Badge shows "False Positive" indicator

To undo: Click Unmark False Positive

Replay Action

  1. Click Replay Action
  2. Review action parameters
  3. Confirm replay in modal
  4. Action is re-submitted for processing

Pagination

Navigate large activity lists:

  • Previous/Next buttons for sequential navigation
  • Page numbers for direct access
  • 10 items per page default display

Upload Agent Logs

Import activity from external sources:

  1. Navigate to Upload Agent Logs card
  2. Select JSON file containing agent actions
  3. File is validated and imported
  4. Activities appear in feed

JSON Format

[
{
"agent_id": "custom-agent",
"action_type": "database_read",
"description": "Query customer records",
"tool_name": "postgres_connector",
"timestamp": "2025-01-15T10:30:00Z"
}
]

Support Integration

Submit issues directly from the activity feed:

  1. Navigate to Need Help? card
  2. Describe your issue
  3. Click Submit Support Request
  4. Confirmation message displayed

API Reference

EndpointMethodDescription
/api/v1/actionsGETList all actions
/api/v1/actions?risk={level}GETFilter by risk
/api/agent-action/false-positive/{id}POSTToggle false positive
/api/v1/actions/upload-jsonPOSTImport actions

Source: ow-ai-backend/routes/actions_v1_routes.py

Data Refresh

  • Auto-refresh: Every 30 seconds
  • Manual refresh: Filter change triggers refresh
  • Real-time: WebSocket updates for critical actions

Best Practices

  1. Filter by status: Focus on pending actions first
  2. Review high-risk items: Prioritize by CVSS score
  3. Use search: Find specific agents or tools quickly
  4. Check compliance mappings: MITRE/NIST references for context
  5. Flag false positives: Improve model accuracy over time

Troubleshooting

Activity not loading

Solution: Check authentication; verify API connectivity.

Filters not working

Solution: Clear search term; reset to "All" options.

CVSS scores missing

Solution: Not all actions have CVSS assessment; risk_score provides internal scoring.

Pagination stuck

Solution: Reset currentPage by changing filters.


Source: AgentActivityFeed.jsx, actions_v1_routes.py