Relevant when…
You want to compare specialist assessments with platform-based issue collaboration.
This is editorial fit guidance based on public positioning. We have not independently tested this provider or confirmed availability.
What the provider documents
- Assessment scope
- AI agent testing · LLM application testing · RAG security · Prompt injection · Tool misuse
- Engagement model
- Hybrid service
- Deliverables
- Platform findings include reproduction steps and remediation guidance.
- Deployment
- Not published
- Pricing
- Not published
- Retesting
- Retesting is part of the described methodology; confirm scope and limits.
Source for documented facts: the provider’s official offering ↗. “Not published” means not recorded in the reviewed source, not that an option is unavailable.
Before you request a proposal
- Which testing level covers our connected agent and tool permissions?
- Are cross-tenant retrieval and indirect injection scenarios included?
- Which evidence exports and data-retention terms apply?
Ask the provider to put authorized systems, exclusions, evidence handling, timing and acceptance criteria in the written scope. Do not send credentials or vulnerability evidence through an initial inquiry.
Sources & verification
- Cobalt — official service information ↗
Supports the summary, service model, scope and recorded engagement facts above. Evidence status: provider-stated.