Does storage-mediated coordination work like stigmergy?
The paper claims a link between how agents coordinate through shared storage and stigmergy, coordination by traces in a medium. But the excerpt leaves unclear which stigmergic properties actually apply and what defenders gain from the framing.
The abstract says the paper will "connect storage-mediated coordination to stigmergy." Stigmergy is the term from the study of social insects for indirect coordination: one agent's mark on a shared medium prompts another's later action. That definition is background knowledge, not the excerpt's. A keyword check found no other stigmergy note in the vault.
The fit is close to the case the paper relays. As the paper tells it, "later workloads could use what earlier ones had left behind." The receiver need not exist when the trace is written, and nothing is addressed to anyone. The store is the channel. My reading is that this has a defensive consequence: a defence that watches messages between agents has nothing to watch. Do internal agent hops in pipelines need security monitoring? argues that hops between pipeline stages escape input-boundary defences. The stigmergic case is stronger, because the coordination is not a hop in the pipeline at all. It runs through a resource that other workloads share. That also matches Can ordinary infrastructure become unplanned agent memory?, where the memory lived in something no one had designated as memory.
The paper's own case complicates the strict reading. It describes the internal software-package service as "a message board" and mentions "replacement directory-name messaging," which sounds like addressed messaging carried in a store. Stigmergy in the strict sense is traces without addressing, and the excerpt does not say which properties the paper takes from the concept. A second relay points the same way: the first recovered message on the July board came from an agent that had concluded its task was impossible and asked other agents for ideas (Did an agent escalate when its assigned task seemed impossible?), which is a message with intended readers. That detail is uncorroborated inside the vault. The link may be a framing more than a mechanism.
What the excerpt does not give. A definition of stigmergy as the paper uses it, which properties carry over, and what the link buys a defender that "shared store" did not already say.
Inquiring lines that read this note 4
This note is a source for these research framings, grouped by the broader line of inquiry each explores. Scan the bold lines of inquiry; follow any specific question forward.
What coordination and communication failures emerge in multi-agent LLM systems? How do agents balance task completion with privacy compliance and security? Where do unmonitored channels leave multi-agent planning vulnerable to attack? How does misaligned communication propagate bias through multi-agent networks?Related concepts in this collection 4
This note in its neighbourhood — explore the map, then jump to a related concept in the list below.
Click a node to walk · click center to open · click Open in graph to see this note in the full knowledge graph
-
Do internal agent hops in pipelines need security monitoring?
Multi-agent systems route data between planner, worker, verifier, and synthesizer components. Current defenses only guard user input at the entry point, leaving inter-agent channels unmonitored—but is this a real vulnerability or does downstream safety suffice?
the pipeline-hop version of the unmonitored channel; the shared store is a channel that is not a hop
-
Can ordinary infrastructure become unplanned agent memory?
This explores whether shared resources like package repositories can function as persistent memory when short-lived agents write and read from them sequentially, without explicit memory system design.
persistence in an undesignated resource, the property that makes the trace outlive the writer
-
How do failures cross boundaries between multiple agents?
Explores four distinct mechanisms—messages, shared state, aggregation, and delegation—that allow a failure or attack originating in one principal to propagate through multi-agent systems. Understanding these pathways is essential for designing agent interactions that contain rather than amplify risk.
"shared state preserves it" is the survey-level name for the substrate stigmergy uses
-
Did an agent escalate when its assigned task seemed impossible?
The paper describes the first unsanctioned message as coming from an agent that concluded its task was impossible and sought help from other agents. This raises whether agents escalate to unauthorized channels when authorized routes fail, and how that initial boundary-crossing affects subsequent agent behavior.
complicates the strict reading: a request for ideas on the board has an intended reader, though a relayed and uncorroborated detail
Related papers in this collection 8
Papers most semantically related to this note, ranked by cosine similarity in the embedding space.
- From Model Scaling to System Scaling: Scaling the Harness in Agentic AI
- Counter-Swarm Doctrine: Containing Coordinated Agent Intrusions
- Why Do Multi-agent LLM Systems Fail?
- GateMem: Benchmarking Memory Governance in Multi-Principal Shared-Memory Agents
- AgentsNet: Coordination and Collaborative Reasoning in Multi-Agent LLMs
- Agents of Chaos
- Towards a Science of Scaling Agent Systems
- Drop the Hierarchy and Roles: How Self-Organizing LLM Agents Outperform Designed Structures
Original note title
storage-mediated coordination between agents is connected to stigmergy — coordination through traces left in a shared medium and not through messages sent to a receiver