Consensus Layer¶

The Consensus layer ensures safety — destructive operations require multi-agent agreement before execution.

Pipeline¶

Destructive operations (file writes, shell commands, HTTP fetches) follow this pipeline:

broadcast → quorum evaluation → red team verification
    → Shapley attribution → trust update → Hebbian learning

A single corrupted agent cannot cause damage.

Components¶

Quorum Engine¶

Collects confidence-weighted votes from agents. Each vote carries:

The agent's decision (approve/reject)
A confidence score (0.0 to 1.0)
The agent's current trust reputation

The quorum passes when the weighted approval exceeds the configured threshold.

Trust Network¶

A Bayesian trust model using Beta distributions. Each agent's reliability is tracked as Beta(α, β):

Success → α increments (trust increases)
Failure → β increments (trust decreases)
Expected trust = α / (α + β)

Trust scores decay slowly over time so agents must maintain performance to keep high trust.

Shapley Attribution¶

After an operation completes, Shapley values determine each agent's marginal contribution to the outcome. This provides fair credit assignment — agents that contributed more to successful outcomes gain more trust.

Escalation¶

A 3-tier failure cascade:

Retry with different agent selection
Escalate to broader quorum
Reject with explanation to the user

Red Team Verification¶

Two red team agents independently re-execute operations to verify results. If a red team agent disagrees with the primary result, the operation is flagged.

A test agent (CorruptedFileReaderAgent) deliberately returns fabricated data to verify that the consensus layer detects and rejects it.

Source Files¶

File	Purpose
`consensus/quorum.py`	Confidence-weighted voting
`consensus/trust.py`	Bayesian Beta(α,β) reputation
`consensus/shapley.py`	Shapley value attribution
`consensus/escalation.py`	3-tier failure cascade