Condor

Agentic AI security scanner — OWASP ASI Top 10

Current release: 1.0.0

Condor scans agentic AI platforms for security vulnerabilities mapped to the OWASP Top 10 for Agentic Applications. It understands the semantics of each platform — enumerating chatflows, tools, vectorstores, and agent surfaces — then runs 10 specialized modules in parallel.

Against a Flowise 1.8.2 instance, Condor finds 9 findings in 3.6 seconds — 3 CRITICAL, 5 HIGH, 1 MEDIUM — all confirmed true positives mapping to published GHSAs.

Part of the CobaltoSec toolkit alongside Corvus (MCP security) and Llamascope (AI infra discovery).

Install

pip install cobaltosec-condor

Quick start

# Scan a Flowise instance
condor scan --url http://localhost:3000 --platform flowise

# JSON output for piping
condor scan --url http://localhost:3000 --platform flowise --format json --stdout | jq '.findings[].severity'

# SARIF output for GitHub Code Scanning
condor scan --url http://localhost:3000 --platform flowise --format sarif

# Authenticated scan
condor scan --url https://letta.internal --platform letta --api-key $LETTA_KEY

# Batch scan from a targets file
condor scan --targets targets.txt --platform generic --concurrency 10 --format json

# Run specific modules only
condor scan --url http://localhost:3000 --platform flowise -m privilege-abuse -m code-execution

Modules — 10 / 10 OWASP ASI

Supported platforms — 16

Output formats

--format table    # Rich terminal table (default)
--format json     # Structured JSON
--format sarif    # SARIF 2.1.0 — GitHub Code Scanning / SAST tooling
--format html     # Self-contained HTML with dark mode and collapsible findings
--format junit    # JUnit XML — Jenkins, GitLab, CircleCI
--format both     # JSON + SARIF written simultaneously
--stdout          # Emit JSON to stdout for piping (suppresses progress bar)

Advanced usage

# Suppress known-good findings with a baseline
condor scan --url http://localhost:3000 --platform flowise --save-baseline baseline.json
condor scan --url http://localhost:3000 --platform flowise --baseline baseline.json

# Route through Burp Suite
condor scan --url http://localhost:3000 --platform flowise \
  --proxy http://127.0.0.1:8080 --insecure

# Only report HIGH and above
condor scan --url http://localhost:3000 --platform flowise --min-severity high

# Load defaults from a config file
condor scan --config condor.yaml

# Scaffold a new module
condor scaffold --name my-check --asi 03

GitHub Actions

- uses: CobaltoSec/condor/.github/actions/condor-scan@v1
  with:
    url: ${{ env.STAGING_URL }}
    platform: flowise
    format: sarif
    fail-on: critical

- uses: github/codeql-action/upload-sarif@v3
  with:
    sarif_file: condor-results/report.sarif

Integrations

# Slack / Teams notifications
condor scan --url http://localhost:3000 --platform flowise \
  --notify-slack $SLACK_WEBHOOK_URL

# DefectDojo export
condor scan --url http://localhost:3000 --platform flowise \
  --defectdojo-url https://defectdojo.internal \
  --defectdojo-token $DD_TOKEN \
  --defectdojo-product "AI Platform Audit"

Plugin system

Third-party modules and platform adapters are auto-discovered via importlib.metadata entry points:

# pyproject.toml of your plugin package
[project.entry-points."condor.modules"]
my-check = "my_package.modules:MyCheckModule"

[project.entry-points."condor.platforms"]
my-platform = "my_package.platforms:MyPlatform"

Compliance

Every finding includes references to ISO/IEC 42001, NIST AI RMF, and EU AI Act. CWE IDs are included in SARIF output for integration with SAST tooling.

Responsible Disclosure

Findings discovered by Condor are reported under the CobaltoSec Responsible Disclosure Policy. Vendors are notified privately via GitHub Security Advisory before any public disclosure.

License

MIT © CobaltoSec