What happens in nature?
Vibrio harveyi's quorum-sensing circuit detects chemical signals called autoinducers and regulates gene expression. Mok, Wingreen and Bassler investigated the joint processing of two signal channels and its relationship to target genes.
From mechanism to protocol
- Test, source and review signals
- Verify independence · check the threshold
- Advance when requirements are satisfied
| In biology | A counterpart to test in your agents |
|---|---|
| Autoinducer channel | Typed validation signal: test, source or review |
| Signal accumulation | Evidence record with expiry and source identity |
| Conditional response | Next stage gated by a predefined rule |
How can you use it in your swarm?
Open the next work stage when independent, valid inputs are ready.
Derive the threshold
Specify prerequisites up front, such as one passing test and a separate source review. Agreement by three agents is not sufficient by itself.
Track signal origin
Attach evidence_id, origin_id and expires_at. Do not count agents copying one source as separate support.
Combine valid inputs
Check every required signal type. Do not override a failing test by collecting more positive votes.
Record the transition
Preserve the signal IDs used in a transition. If conditions fail, report an explicit waiting reason or timeout.
Start with an experiment
Compare a fixed majority threshold with a gate checking source, age and type. Include duplicate and stale signals.
- What happens if you remove the mechanism?
- Disable source deduplication. Pass the same wrong report through five agents and measure false gate openings.
- Primary failure risk
- Duplicate reports can create false sufficiency. An old passing test does not validate a new artifact version; bind signals to versions.
Evidence and related studies
Read biological evidence and agent research separately. The engineering interpretations below are SWI synthesis.
Vibrio harveyi quorum sensing: a coincidence detector for two autoinducers controls gene expression
V. harveyi integrates signals from two autoinducer channels in a circuit regulating gene expression.
What can I use? Interpretation, limits and provenance
SWI engineering interpretation
Combine multiple independent signals before triggering an expensive action.
Limitation
Chemical concentration is not a confidence score; copied agent reports are not independent evidence.
Source record
Kenny C. Mok, Ned S. Wingreen, Bonnie L. Bassler
Publication: 2003 · exact day not verified
Abstract review · Checked: 2026-09-06
Why Do Multi-Agent LLM Systems Fail?
Version 3 presents a dataset of over 1,600 traces and 14 failure modes grouped into design, inter-agent misalignment and verification.
What can I use? Interpretation, limits and provenance
SWI engineering interpretation
Log failure causes such as wrong tasks, broken handoffs and premature termination alongside outcomes.
Limitation
The taxonomy is a substantial starting point, not a claim to cover every failure in all systems.
Source record
Mert Cemri et al.
Publication: 2025-03-17
Reviewed revision: 2025-10-26
Abstract review · Checked: 2026-09-06
Towards a Science of Scaling Agent Systems
Version 3 compares 260 configurations across six benchmarks; coordination benefits depend on task structure and communication overhead.
What can I use? Interpretation, limits and provenance
SWI engineering interpretation
Compare single-agent, independent parallel and coordinated teams at the same aggregate budget.
Limitation
Reported counts changed between versions; this entry uses v3. Findings are not a universal scaling law.
Source record
Yubin Kim et al.
Publication: 2025-12-09
Reviewed revision: 2026-04-08
Abstract review · Checked: 2026-09-06