Build on the Security Swarm Platform.
Launch swarms, pull proven findings, and retest fixes from your own code.
One API for the full loop.
Use the REST API from services and pipelines, and the MCP server from agents.
Launch swarms
Start black-box or white-box assessments at Balanced, Deep, or Ultra depth.
Pull proven findings
Every finding includes its evidence, impact, reproduction steps, and remediation advice.
Retest fixes
Run the original attack again after a fix ships. Only a confirmed fix resolves the finding.
Measure coverage
See which security scenarios are covered, failing, or incomplete for each target.
Generate reports
Create stakeholder-ready reports and download them.
Validate and schedule from agents
Submit outside reports to Validator and schedule recurring runs over MCP.
From first request to verified fix.
Start a black-box or white-box assessment at Balanced, Deep, or Ultra depth. Pass a primary and a secondary credential to test whether one account can reach another account's data.
# Launch an authenticated, cross-account assessment
curl -X POST https://api.mindfort.app/v1/assessments/run \
-H "Authorization: Bearer $MINDFORT_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"target_id": "'"$TARGET_ID"'",
"assessment_method": "ultra",
"target_credential_id": "'"$PRIMARY_CREDENTIAL_ID"'",
"secondary_target_credential_id": "'"$SECONDARY_CREDENTIAL_ID"'"
}'
# Track each stage until the run completes
curl https://api.mindfort.app/v1/assessments/$ASSESSMENT_ID \
-H "Authorization: Bearer $MINDFORT_API_KEY"
Put the swarm where you work.
The engine that runs thousands of agents for MindFort customers can run behind your product too.
Inside your platform
Run a swarm against each authorized deployment and show proven findings in your own product.
In every pipeline
Trigger a swarm on every push or release, and hold the release when critical findings come back.
In your agents
Give Cursor, Claude Code, Codex, or your own agents direct access to findings and retests over MCP.
Programmatic, not unbounded.
Runs started from code follow the same scope, guardrails, and monitoring as runs started from the dashboard.
How MindFort keeps agents in scope