Scanners produce more findings than anyone can act on, and the two questions that matter are 'is this the same bug in fifteen places' and 'does this code do anything important'. Most tools answer neither. Having both as callable tools means an agent can group and prioritise rather than reading down a list. Status changes being audit-logged is the right complement — triage by agent should leave a trail.
CybeDefend's endpoint, published in two regions. Eighteen tools, and the structure is one list-and-get pair per scan type — six scan types, plus the tools that put a finding in context.
- List and get pairs for SAST, SCA, IaC, CI/CD, secret and container vulnerabilities
- list_sca_packages — the dependency inventory behind the SCA findings
- get_similar_vulnerabilities — other instances of the same problem
- get_business_logic_context — what the affected code is actually for
- get_project_overview, get_branches, update_vulnerability — the project, its branches, and triage
OAuth 2.1 with dynamic client registration and browser sign-in. Region is chosen at signup — the EU deployment runs on Scaleway in Paris under GDPR, NIS2 and DORA. Free start: 50 AI credits and 10 scans with no card; every plan includes all scan types.
Paste a URL, then authorize — add the endpoint to your client, then approve the OAuth consent screen