[Hackathon] stanford-ml-phd: EigenTrust plugin for the trust layer

README.md

@@ -227,7 +227,7 @@ resolved by name via entry points or a built-in default.
 |  3 | Identity      | `Identity`      | `did_key` (deterministic public-key signatures for simulation; not Ed25519) |
 |  4 | Registry      | `Registry`      | `in_memory` (dict lookup, no persistence) |
 |  5 | Auth          | `Auth`          | `jwt` (HMAC-SHA256 token; not RFC JWT) |
-|  6 | Trust         | `Trust`         | `score_average` (running mean reputation; no Sybil resistance) |
+|  6 | Trust         | `Trust`         | `score_average` (running mean reputation; no Sybil resistance). Also bundled: `eigentrust` (transitive, Sybil-resistant). |
 |  7 | Payments      | `Payments`      | `prepaid_credits` (in-memory ledger) |
 |  8 | Coordination  | `Coordination`  | `contract_net` (FIPA: propose · bid · resolve · commit) |
 |  9 | Negotiation   | `Negotiation`   | `alternating_offers` (Rubinstein, with patience discount) |

docs/layers/trust.md

@@ -15,20 +15,26 @@ class Trust(Protocol):
 
 Full definition: [`nest_core/layers/trust.py`](../../packages/nest-core/nest_core/layers/trust.py).
 
-## Default plugin
+## Built-in plugins
 
-`score_average` — running mean of feedback scores. No Sybil resistance,
-no decay, no stake economics.
+| name | summary | Sybil-resistant? |
+|---|---|---|
+| `score_average` (default) | Running mean of feedback scores. | No |
+| `eigentrust` | Transitive reputation via the EigenTrust eigenvector (Kamvar et al., WWW '03). Weighs each report by the reporter's current trust, with a teleport to a configurable pre-trusted seed set. | Yes |
 
-Source: [`nest_plugins_reference/trust/score_average.py`](../../packages/nest-plugins-reference/nest_plugins_reference/trust/score_average.py).
+Sources:
+- [`nest_plugins_reference/trust/score_average.py`](../../packages/nest-plugins-reference/nest_plugins_reference/trust/score_average.py)
+- [`nest_plugins_reference/trust/eigentrust.py`](../../packages/nest-plugins-reference/nest_plugins_reference/trust/eigentrust.py)
 
 The `reputation` scenario exercises this layer — 16 honest + 4
 malicious + 1 observer that samples cheat reports probabilistically.
+Swap `trust: score_average` for `trust: eigentrust` in the scenario
+YAML to compare adversarial behaviour under the two ranking rules.
 
 ## Writing your own
 
 See [`writing-a-plugin.md`](../writing-a-plugin.md). Register under
 entry point group `nest.plugins.trust`.
 
-Good fits to test here: EigenTrust-style transitive reputation,
-proof-of-stake reputation, decaying scores, attestation graphs.
+Good fits to test here: proof-of-stake reputation, decaying / time-
+weighted scores, attestation graphs, learned reputation models.

packages/nest-core/nest_core/plugins.py

@@ -23,6 +23,7 @@
     ("registry", "in_memory"): f"{_REF}.registry.in_memory:InMemoryRegistry",
     ("auth", "jwt"): f"{_REF}.auth.jwt_auth:JwtAuth",
     ("trust", "score_average"): f"{_REF}.trust.score_average:ScoreAverageTrust",
+    ("trust", "eigentrust"): f"{_REF}.trust.eigentrust:EigenTrust",
     ("payments", "prepaid_credits"): f"{_REF}.payments.prepaid_credits:PrepaidCredits",
     ("coordination", "contract_net"): f"{_REF}.coordination.contract_net:ContractNet",
     ("negotiation", "alternating_offers"): (