Research trends

Research trends in multi-agent security.

A quick view of which topic areas are currently leading the stored corpus, which ones are building momentum, and which ones are still earlier in the research cycle.

Trend board

Current topic momentum

This is a directional read on the stored corpus rather than a formal forecast. Larger tracks indicate stronger concentration in the current article set.

Leading signal

Governance and Policy

925

Strong current attention relative to the rest of the tracked research landscape.

Leading signal

Benchmarks and Evaluation

599

Strong current attention relative to the rest of the tracked research landscape.

Leading signal

Orchestration Risk

560

Strong current attention relative to the rest of the tracked research landscape.

Building momentum

Agent-to-Agent Communication

532

Active and visible, with room to become a more central thread in the literature.

Building momentum

Trust and Identity

497

Active and visible, with room to become a more central thread in the literature.

Building momentum

Prompt Injection

289

Active and visible, with room to become a more central thread in the literature.

Taxonomy detail

Current topic breakdown

Click a row to open the papers currently mapped into that multi-agent security area.

Area	Papers	Trend signal	Interpretation
Governance and Policy	925	High concentration	Leading current theme with the strongest concentration in the stored corpus.
Benchmarks and Evaluation	599	High concentration	Leading current theme with the strongest concentration in the stored corpus.
Orchestration Risk	560	High concentration	Leading current theme with the strongest concentration in the stored corpus.
Agent-to-Agent Communication	532	Emerging / moderate	Visible and active, but not yet as dominant as the leading categories.
Trust and Identity	497	Emerging / moderate	Visible and active, but not yet as dominant as the leading categories.
Prompt Injection	289	Emerging / moderate	Visible and active, but not yet as dominant as the leading categories.
Memory Poisoning	135	Emerging / moderate	Visible and active, but not yet as dominant as the leading categories.