NERV Agent turns a request into an auditable sequence of actions. It can combine several capabilities in one investigation while keeping evidence, decisions and generated material tied to the active case.
Core capabilities
- Identify people, organisations, domains and other entities, then ask the analyst to resolve ambiguity before deeper research.
- Search public sources, read pages and documents, analyse images and organise attributable evidence.
- Build, clean and export an intelligence graph; place relevant facts on a map or timeline.
- Draft structured reports and reusable views from the investigation.
- Run authorised awareness simulations, privacy-preserving collaboration sessions and analyst-made automations.
- Analyse financial, network and technical material when the required capability is available and authorised.
Typical flow
- State objective
- Resolve identity
- Approve scope
- Run capabilities
- Review evidence
- The analyst states the objective and opens the relevant case.
- The agent identifies the task, available context and required approval boundaries.
- It proposes questions or candidate identities when a decision belongs to the analyst.
- It performs the selected actions and records outputs in the case.
- The analyst reviews evidence, corrects assumptions and decides what becomes part of the record.
Control remains visible
Capabilities do not bypass permissions. External actions, sensitive analysis and operations that affect third parties require the appropriate configuration and authorisation. The agent can judge relevance and coherence; provenance, policy and workspace boundaries remain enforced by the system.