Protocols +
plugins.
Every pull request on projnanda/nandatown, synced from GitHub. Explore open, merged, and closed contributions across twelve protocol layers, with work beyond the stack collected under Other.
358
111
73
174
12/12
Transport
How bytes move between agents.
Communication
Message framing and request/response semantics.
Identity
Sign and verify per-agent payloads.
Registry
Publish and discover agent cards.
Auth
Issue, verify, and revoke capability tokens.
Trust
Reputation scores, attestations, reports.
Payments
Quote, pay, verify, refund.
Coordination
Group decisions and task allocation.
Negotiation
Bilateral bargaining.
Memory
Shared key-value with subscribe and CAS.
Privacy
Encryption and zero-knowledge proofs.
Data Facts
Dataset publish, fetch, and ACL.
Other
Builds that reach beyond the twelve layers.
All pull requests
Merged work,
open reviews and history.
[Fellowship Challenge] registry: index.owned.v1, forbidding a forged card from erasing an honest listing
Submission for the Hexaware x Project NANDA Fellowship Technical Challenge
[Fellowship Challenge] Track: quote-amnesia-restart, a seller that forgets work it already answered
Submission for the Hexaware x Project NANDA Fellowship Technical Challenge
[Fellowship Challenge] trust: reputation.capped.v1, a per-observer influence cap
trust/reputation.v1 sums +1/-1 per receipt with no bound on how much any single observer can move a score. reputation_consistent is explicit that it checks the recorded claim, not whether the reporter told the truth or held authority.
[Fellowship Challenge] Path profile for a reused order id with changed terms
Submission for the Hexaware x Project NANDA Fellowship Technical Challenge
fix(dashboard): check a submitted source_url is an http(s) URL again
Flagged by an automated security review of main, and confirmed by hand.
fix(identity): let concurrent runs share a keystore without breaking attestations
Follow-up found while reviewing the merged 00a013b, prioritised by Codex.
fix(replay): step through Path bundles instead of crashing
Follow-up found while reviewing the merged 00a013b, raised by Codex.
fix(pulse): record a malformed redirect as the answer the server gave
Lower-priority follow-up to the review of merged 00a013b.
fix: use credentials written into an endpoint URL, never record them
Follow-up to the review of merged 00a013b: URL confidentiality.
docs: say which historical bundles replay, and which cannot
Lower-priority follow-up to the review of merged 00a013b: statements about
fix(evaluator): 0.5.0 binds duplicate delivery to the injected offer and lets only a terminal buyer verdict decide
Follow-up to the review of merged 00a013b. Two of its findings change what the
fix(track): report an application whose acknowledgement was fenced
Follow-up to #264–#277, from the same correctness audit of main at df0b5f1.
fix(coordinator): refuse JSON values the run cannot export
Follow-up to #264–#272, from the same correctness audit of main at df0b5f1.
fix(bundle): read JSON Lines records at LF only
Follow-up to #264–#272, from the same correctness audit of main at df0b5f1.
fix(cli): stop Track processes when the CLI receives SIGTERM
Follow-up to #264–#272, from the same correctness audit of main at df0b5f1.
fix(runner): refuse a Track timeout too short for Town's counterparts
Follow-up to #264–#272, from the same correctness audit of main at df0b5f1.
test(participants): wait for the seller's ack before asserting it
Follow-up to #264–#272, from the same correctness audit of main at df0b5f1.
fix(pulse): refuse a reused --target name
One of nine independent PRs from a correctness audit of main at df0b5f1. Each merges cleanly with the others in any order; together they pass 908 tests locally on Python 3.11 and 3.14 (CI covers 3.11 and 3.12).
fix(path): fail resolution on a malformed local index
One of nine independent PRs from a correctness audit of main at df0b5f1. Each merges cleanly with the others in any order; together they pass 908 tests locally on Python 3.11 and 3.14 (CI covers 3.11 and 3.12).
fix(path): record undecodable terminal output as the subject's
One of nine independent PRs from a correctness audit of main at df0b5f1. Each merges cleanly with the others in any order; together they pass 908 tests locally on Python 3.11 and 3.14 (CI covers 3.11 and 3.12).
fix(cli): print join credentials for run --agent ROLE=external
One of nine independent PRs from a correctness audit of main at df0b5f1. Each merges cleanly with the others in any order; together they pass 908 tests locally on Python 3.11 and 3.14 (CI covers 3.11 and 3.12).
fix(runner): wait for an external buyer's verdict before finishing
One of nine independent PRs from a correctness audit of main at df0b5f1. Each merges cleanly with the others in any order; together they pass 908 tests locally on Python 3.11 and 3.14 (CI covers 3.11 and 3.12).
fix(pulse): attribute availability to the endpoint actually probed
One of nine independent PRs from a correctness audit of main at df0b5f1. Each merges cleanly with the others in any order; together they pass 908 tests locally on Python 3.11 and 3.14 (CI covers 3.11 and 3.12).
fix(receipt): refuse receipts over bundles that fail verification
One of nine independent PRs from a correctness audit of main at df0b5f1. Each merges cleanly with the others in any order; together they pass 908 tests locally on Python 3.11 and 3.14 (CI covers 3.11 and 3.12).
fix(path): keep the tested endpoint and its rerun exact
One of nine independent PRs from a correctness audit of main at df0b5f1. Each merges cleanly with the others in any order; together they pass 908 tests locally on Python 3.11 and 3.14 (CI covers 3.11 and 3.12).
fix(evaluator): require boolean acknowledgement flags and judge every accepted request
Two pre-existing evaluator gaps, reported against #264 and confirmed to
fix(evaluator): fail Track runs with duplicate or uncorrelated quote responses
One of nine independent PRs from a correctness audit of main at df0b5f1. Each merges cleanly with the others in any order; together they pass 908 tests locally on Python 3.11 and 3.14 (CI covers 3.11 and 3.12).
docs(town): align the current manual, trace viewer and build checks
Bring the public entry points into agreement with the rebuilt Town code and its actual evidence boundaries:
fix(packaging): ship runtime assets and verify built distributions
- Regenerate stale exported record schemas and test them against the live models.
fix(catalog): bound submissions and commit history atomically
- Bound streamed JSON request bodies and individual catalog fields before any database work, with shared validation for the API and submission form.
fix(viewers): safely render evidence and rank only verified bundles
- Render participant-controlled viewer values as inert text, escape the HTML title/script-data boundary, constrain styles, and support the actual Path profile and empty timelines.
fix(imports): preserve pinned snapshots and disclose incomplete imports
- Refuse silent replacement of an on-ramp snapshot and validate the full pinned release reference on an idempotent re-import.
fix(runtime): serialize mailbox work and settle participant processes
- Serialize mailbox claim/ack/reoffer decisions with their writes and evidence events; finish is atomic/idempotent and terminal runs reject subsequent mutations.
fix(evidence): bind receipt claims and version strict Path evaluation
A valid signature alone was too easily confused with agreement between a receipt's claims and its bundle. Incomplete/malformed bundle manifests and replay ambiguities could also undermine downstream verification.
Bind marketplace settlements to capped negotiations
- Enforce the stock marketplace buyer's configured unit-price cap before an order or escrow hold can be created.
Fail closed on invalid fault-rule numbers
- Validate every current fault-rule numeric field before a run starts.
Reject duplicate bids in sealed-bid auctions
- Make the current advertised sealed-bid auction accept at most one bid from each bidder before award.
Stop fetching untrusted catalog URLs on submission
- Stop the catalog form from fetching user-submitted URL and GitHub sources.
Reject malformed ledger amounts before mutation
- Enforce the declared integer-cent domain at every current ledger entry boundary.
Reject malformed run grant timestamps without changing signed payloads
Reject malformed or non-finite timestamps when verifying current Town run grants, without changing signed payloads or adding permission policy.
Make message acceptance atomic and bind the complete envelope
Make current Town's durable message acceptance atomic and bind replay to the complete message envelope.
Add optional quote-intent profile for cheaper wrong-item failures
Add the optional a2a-quote-intent@0.1 profile to the existing Path runner. A quote within the price ceiling is not enough if it describes the wrong item.
test: require run-bound session proof at grant join
Preserves the absent/wrong-presenter test requirement from legacy #156 (@mwarsss) on current Town Run Grants. The production check already exists; this PR adds seven regressions only, with no runtime, permission or schema change.
fix(catalog): guard form submission re-entry
Rescues the client-side re-entry guard from @stanleyoz's legacy #204 onto current main. The retained SkillMD form could queue a second action when submit events arrived before the pending render or while the first action was outstanding.
fix: validate catalog sources and label unchecked links
Preserve the useful current-catalog requirements from #130 by @nonggde and the review by @dhve, without adding another automatic remote probe.
fix: bind Lab auction settlement to its task and payee
Current Lab auction settlement checks the payer and amount but not the payee or task. A real local run with the ledger transfer redirected to a losing bidder still reports passed, even though the auctioneer receives zero.
fix: verify correlated majority evidence in Lab consensus
Fix a false-positive in the current Lab consensus evaluator. It previously counted delivered prepare_ack events and entries in the proposer’s acks list without independently establishing distinct eligible voters or binding the…
test: preserve grant key-binding regressions from legacy #172
Retain the key-substitution regression requirement from #172 by @theCodeForgerHQ against current Run Grants.
fix(lab): verify receipt-backed reputation arithmetic
The marketplace reputation stage previously accepted two updates ending at score 2, even if receipt events were absent, a receipt named a different seller, or the first score was 999.
fix(path): require exact fulfillment request correlation
Require a fulfillment's returned request_id to match the exact, nonempty order identifier Town issued and recorded as the event subject.
fix(runner): reject unknown agent override roles
Reject unsupported role keys in both harnesses and external before starting a Track run.
fix(receipt): verify evidence before rendering Town Proof
Require successful bundle verification before rendering Town Proof or automatically creating its receipt.
fix(bundle): verify the declared bundle fingerprint
Verify that the manifest's declared bundle_fingerprint matches the canonical fingerprint of its file-hash map before accepting a bundle.
fix(lab): disclose untested scope of imported scenarios
Imported legacy scenarios run Town reference agents, not their original plugins or validators. Previously a declared network_partition disappeared silently, and a generic adapted exchange could pass without any result saying that the…
fix(a2a): bound streamed responses and version transport policy
Bound external AgentCard and native A2A JSON responses before parsing, and record the effective transport policy in each Path run.
fix(lab): require scenario coverage before a passing verdict
A registered Lab validator can return no scenario checks and still receive a passing verdict from generic checks. This change makes missing scenario coverage explicit before those generic checks are appended.
fix(path): preserve the explicit profile in rerun commands
Generated Path rerun commands used --profile, but the already-running-agent CLI consumes --path-profile. A rerun therefore used the current default instead of the profile recorded in the evidence bundle.
feat(agent-test): test an existing OpenClaw agent from the CLI
The headline workflow should be one clear command for a terminal-comfortable developer, not a private bridge or hand-edited adapter.
feat(agent-test): run managed runtimes through Town
A prepared runtime still needs lifecycle-safe orchestration through the shared loopback adapter and Town evaluator without coupling the engine to OpenClaw internals.
feat(agent-test): add a bounded OpenClaw connector
Testing an existing OpenClaw agent needs a bounded connector that checks the behavior Town relies on without locking users to one release or guessing across incompatible routes, retries, or output envelopes.
refactor(agent-test): share the local adapter server
The managed-runtime path needs the same strict authenticated loopback server as the public reference adapter without duplicating or drifting its wire behavior.
feat(agent-test): add the local-adapter CLI
Developers need a documented command to attach a strict local adapter, run the Town test, and consume stable human or machine-readable results.
feat(agent-test): evaluate external agents in Town
A successful adapter exchange is not enough: Town must drive the participant through a real deterministic workflow and emit honest, stage-specific evidence instead of a single opaque score.
feat(agent-test): add a hardened loopback HTTP driver
Town needs a small local adapter transport with strict authentication, bounds, timeouts, and failure attribution before external agents can safely participate in a test run.
feat(agent-test): define version 1 contracts and profiles
External-agent tests need a strict, versioned contract whose wire bytes, profile identity, result semantics, and extension boundary are reviewable independently of transport and CLI code.
feat(town): add deterministic capability testing support
Town needs a deterministic, real execution seam before an external agent can be evaluated through the simulator without coupling generic scenario code to the agent-test feature.
Fix/enforce grant permissions
Merging branch with added visualization.
Enforce Run Grant permissions on join and every mailbox action
nandatown run --identity joins each role through a Run Grant: a controller-signed authorization for one disposable session key, for one run, with named permissions.
Add Agentic Commerce Hackathon page
Adds a page promoting the Agentic Commerce Hackathon (Devfolio, July 31 - Aug 2 2026), where Project NANDA is a partner sponsor.
Trust/Payments layer: verifiable capsule receipts (capsule-emit-nanda) — rebuild addressing #177
Re-opens the work from #54 → #177 (closed by @Skyrider3), rebuilt to address every
Trust layer: verifiable capsule receipts with CCF write-receipt anchoring
Re-opens the work from #54 → #177 (closed by @Skyrider3), rebuilt with a fundamentally
Nanda Town UI: audience-choice voting, live PR gallery, and site cleanup
Brings the live nandatown.projectnanda.org UI onto the repo. All commits authored by dhve. Changes are confined to apps/nest-dashboard/ (no packages/ or scenarios/ touched), so this composes with the merged hackathon PRs already on main.
Revert "[Hackathon] distributed-sys-engg: deterministic negotiation session ids (ADR-004 sibling fix)"
Reverts projnanda/nandatown#184
distributed-sys-engg: deterministic negotiation session ids (ADR-004 sibling fix)
Persona: distributed-sys-engg — reliability/determinism focus.
calyirex: validators must not certify corrupted ground truth (3 fixes: auction shill, voting inflation, revocation laundering)
Three validator fixes, one defect class: a validator that reconstructs 'ground truth' from attacker-influenceable data ends up certifying the very corruption it exists to catch.
replay-safe: comms replay-attack resistance on top of authenticated
I picked #01 — versioned message schemas
[hackathon phase 1] fix: return defensive copy from CidFacts.fetch()
Fixes a bug where CidFacts.fetch() returned the mutable internal object instead of a copy, which allowed the caller to unintentionally tamper with the registry content.
dashrath: agent_receipts — sever collusion by evidence shape, not size (closes #97)
Credit to @SwasthikaDev's issue #97 for identifying the flaw.
dhritimandas: capability-conformance registry gate
AgentCard.capabilities (nest_core/types.py:82) is a bare, self-asserted list[str]. Registry.lookup() does string matching against this list.
Hackathon/parsa noah mesh revocable auth
delegatable (#138) solved delegation, but its revocation set is one in-process
Redeem_Grimm: authenticate failure-detector heartbeats against a forgery attack
This is a direct follow-up to the failure-detector layer merged in #46. In the review on merge, @Skyrider3 flagged two gaps, and both are worth fixing:
ckeda-security: delegatable capability tokens with cascading revocation
Branch: hackathon/ckeda-delegatable-auth · Problem 04
bogacsmz: byzantine-resistant gossip registry — signed-equivocation quarantine + eclipse resistance
Persona for this review: trust/honesty auditor. I built the adversarial
libre: attested-peering trust plugin
Persona: libre — a self-sovereignty / decentralized-infrastructure engineer
moltpass.dev — bonded_trust — a Sybil-resistant trust root with a pluggable scarcity anchor
[Hackathon] Moltpass.dev — bonded_trust — a Sybil-resistant trust root with a pluggable scarcity anchor
validators: fix auction_winner_highest trusting the announced amount over real bids
Fixes validate_auction_winner_highest returning PASS for an auction where a lower bidder was awarded, when the won: message announces a winning amount inflated past every real bid.
hackathon/sam04-trust-gated-privacy
The merged hybrid_x25519 plugin (#28) answers *who* can read with a static, all-or-nothing audience — it is deliberately trust-blind. Meanwhile the trust layer computes reputation scores that no privacy plugin consumes.
edge: address EMPIC escrow review feedback
Follow-up to #41 addressing Skyrider3's review feedback on EMPIC Step 1 escrow behavior.
protocol-security-engineer: tamper-evident comms that resist version downgrade
A new authenticated comms plugin. It's versioned plus an HMAC-SHA256 tag over
stellarminds-ai: AAE permit gate — pre-action authorization with signed denial receipts
Layer: trust · Persona: stellarminds-ai
trust: parc — selective disclosure via Merkle inclusion proofs (stacked on #63)
Selective disclosure for trust: parc credentials: a holder reveals only chosen receipts from their ledger, each with a Merkle inclusion proof against the credential's signed behavioral_merkle_root — a verifier confirms every revealed…
stellarminds-ai: PARC — portable reputation credentials with a recomputing admission gate
Reputation in NEST dies with the run. Every trust plugin holds an in-memory ledger; nothing exports it, carries it, or verifies it in another trust domain — an agent that migrates starts from zero, and a domain that wants to admit a…
Redeem_Grimm: failure-detector layer with a phi-accrual liveness oracle
NANDA Town models failure by injection. message_drop, a Byzantine fraction, and partitions perturb the network, and the validators check that a protocol stays correct in spite of them.
edge: EMPIC escrow payments plugin and adversarial validators
Note: This Step 1 PR is linked to this Step 2 skill: http://edgemicropayments.ddns.net:8099/skill.md
kavya-escrow-payments: conditional, arbitrated escrow plugin for the payments layer
The payments layer currently has two plugins: prepaid_credits (a 122-line debit/credit ledger) and the merged PR #21 streaming plugin (per-tick streaming with mid-stream cancel).
swamx: Partition-tolerant BFT coordination with HotStuff
[Hackathon] swamx: Partition-tolerant BFT coordination with HotStuff view-change
data-engineer: content-addressed datafacts with provenance and signed freshness
[08 — content-addressed datasets with provenance chains and freshness proofs](docs/hackathon/problems/08-datafacts-content-addressed.md).
simon-mechanism-design: Pareto-seeking multi-attribute negotiation with a dominance validator
The reference alternating_offers plugin collapses multi-attribute bargaining into a one-dimensional price fight: its respond only looks at terms.price and never reads conditions['deadline_days'].
[NandaHack] privacy: hybrid_x25519 — real hybrid encryption, selective disclosure & broadcast revocation
Persona: applied-cryptography / security engineer. The code is
stellarminds-ai: Receipt-corroborated reputation with collusion-ring severance
Layer: trust · Persona: stellarminds-ai
stellarminds-ai: Real Ed25519 identity with key rotation and as-of verification
Problem: 05-identity-ed25519-key-rotation · Persona: stellarminds-ai · Layer: identity
bori7-onchain-registry-author: gossip registry plugin with partition-honest eventual consistency
[#06 registry-gossip-eventual-consistency](https://github.com/projnanda/nandatown/blob/main/docs/hackathon/problems/06-registry-gossip-eventual-consistency.md).
[NANDAHack] Improve memory layer: LWW-Register CRDT plugin with convergence validators
Adds a state-based last-writer-wins register CRDT (lww_register) for the memory layer, fixing the blackboard's order-dependent, silently-diverging writes.
feat(payments): streaming per-tick payments with mid-stream cancellation
Problem #3: [Streaming pay-per-second payments with mid-stream cancellation](https://github.com/projnanda/nandatown/blob/main/docs/hackathon/problems/03-payments-streaming-x402.md)
wire-compat: versioned comms layer with forward/backward …
…compat
[Platform] Hackathon marketplace UI: /hackathon section with submissions, layers, authors
Extends the existing apps/nest-dashboard Next.js 16 app (App Router, React 19, Tailwind 4). Same cream / ink / rust palette, Instrument Serif / Geist Sans / Geist Mono, same paper-texture, eyebrow, btn-primary utilities — no new design…
[Platform] Research harness for A/B hackathon experiments
Ships scripts/harness/ — multi-condition, reproducible experiment infrastructure to turn the hackathon into a publishable AI benchmark.
[Platform] Judge panel + scoreboard for hackathon PRs
The NEST hackathon is month-long and aimed at thousands of participants. Agent and human submissions both need to be scored on the same rubric, mechanically, and at scale.
[Platform] Open problems + charter + judging doc
Ships the participant-facing scaffolding for the NEST hackathon:
[Platform] CI hygiene: Makefile, pre-commit, feedback bot, Definition of Done
Closes the gap where hackathon contributors run ruff check + pytest, call that "the tests", and ship PRs that fail CI on ruff format --check and/or pyright.
[Platform] Integration v2
Rebuild post-schema-recon + post-live-judge. PR #17 was originally cut before the schema-recon and live-judge commits landed on platform/judge-panel and platform/marketplace-ui.
LOOM HW9: agent interface, portal 0.2.1, cold-start tests (builds on #362)
Team LOOM (Qingyun Liu, Jiayi Li, Jason Xuan). This branch starts from the HW8 integration in #362 and adds the HW9 work under examples/loom/. It does not change Town's own code.
auth: hmac.bound.v1, so a signed message cannot be replayed to someone else as something else
hmac.v1 signs the message body alone. Whoever receives a signed message can put the same body and signature in a new envelope, under the original sender's name, addressed to anyone and labelled as any kind. It verifies.
registry_snapshot: refuse a former controller whose signatures are valid
A signature proves that a key signed. It cannot tell you whether that key still speaks for the identity it claims. If an external registry moved control of an identity from key K1 to key K2, K1 still signs perfectly, and only the…
Graveyard: record what a fail-closed outside dependency costs under a lease
The Agentic Graveyard is a pre-mortem service for agents: a record of failures, each stored with the check that would have prevented it, served over MCP. This PR does not add the service.
Signed Answers: add outlet source-package Lab integration with duplicate-delivery and altered-package tests
Signed Answers from News Outlets for AI Agents lets a news outlet run its own agent. A reader’s agent asks it about the outlet’s reporting over A2A. The outlet’s agent checks a subscriber credential.
Mine Context Layer: add MCP shared-context integration with permission and degraded-source tests
Mine Context Layer is an MCP server that holds the shared context of one underground-mine shift: plan, fleet, operating rules, events with provenance (source, validity, confidence, link to the original record), proposals, and a human…
SKOPOS: Guardian for Golden Dome
Summary
LOOM: add task-scoped health-context access integration with overreach, expiry and replay tests
Supersedes #329 (same integration, rebased onto current main, plus the replay fix described under Results).
Referral: add checked-work trust layer and fabricated-citation test
[Referral](https://github.com/louis1mx/referral) keeps verified track records for agents that hire agents. Before a job, the worker locks a forecast.
Rebbi: token-lending MCP integration
Rebbi lends spare token budget between the AI agents of one fleet. An agent about to run out mid-job calls post_handoff over MCP; Rebbi books a loan from an agent with budget above its reserve (LEND), or refuses with each lender's reason…
Cerata: add legal-check Track integration with fault and fail-closed tests
Cerata is a hosted legal check for AI agents. Before an agent calls, texts, emails, lists a home or posts a job, it tells Cerata what it is about to do; Cerata answers clear, flag, hold or block, with a signed receipt anyone can verify…
gitmem: Track seller with git-backed memory, plus hand-edited-memory and crash-restart tests
gitmem is a memory layer for AI agents that stores each remembered fact as a file in a git repository, so every change to an agent's memory has a history: what changed, when, who made the change and why.
VeriFlex: integrate Bona energy-flexibility trading with NandaTown and test lost-response handling
VeriFlex is our course project, using the Bona trust engine. Bona is the MCP trust service used by VeriFlex. It serves AI agents that trade energy flexibility.
Agent Trust Layer: add atl.v1 payments layer, real-agent buyer and forged-payee test
Agent Trust Layer (ATL) is a payment guard for AI agents. The agent never holds payment credentials. It submits
Common Ground: three Claude Code agents plan a meetup through the room, with duplicate and dropout tests
Common Ground is a neutral room where people's personal agents plan a real-world meetup for them. Each agent reads its person's private calendar and profile through that person's own MCP server, shares only typed fields with the room (free…
KnownTruth: add qualification-gate buyer, MCP-edge candidate seller and refusal test
KnownTruth builds a test environment from an organization's records with the correct answers planted by construction, runs candidate agents on it, grades them against the planted truth, and serves the results to other agents over MCP.
Let verify load plugin files, and say when a validator is missing
A Lab scenario can name a validator that a plugin file registers (nandatown run scenario.yaml --plugin my_plugin.py). The run works and writes a bundle. nandatown verify on that bundle then fails with:
Attest: add Know-Your-Agent gated seller, signing buyer and duplicate-delivery test
Attest is a Know-Your-Agent gate: before an agent acts on another agent's request it checks who deployed the sender, what the sender may do, and whether that is still valid, and returns allow, refuse, escalate or unknown with a signed…
Pearl Street: metered peak-reduction market in the Lab, with an overpromising provider
Runs Pearl Street, a flexibility exchange for energy agents, end to end in the Town Lab. A utility posts a 30 MW peak-reduction need. Five sites register, find the exchange in Town's index and offer kW.
MindVault: shared-memory seller adapter with crash-restart check
MindVault is shared memory for a team of agents. Every note has an access level (L0 public up to L2 finance), an agent only reads notes at or below its own level, and a note an agent writes is labelled with the highest level that agent has…
Locard: add coverage-gated vendor integration with suspension, tamper and fail-closed tests
Locard is a risk market for AI agents: it scores an agent from its tool manifest and trace history, prices and binds cover, and issues a signed policy token.
Lost Link: test the end-to-end return lifecycle and approval boundaries
Lost Link is a shared lost-and-found service for personal agents. Add a focused Town Lab integration that exercises the separately running project from found/lost registration through matching, private claims, clarification,…
Fervo: add agent-contribution adapter and redelivery-idempotency test
Fervo is repo management for agent contributors. It has a server, a runner and a CLI that let many coding agents improve one Git repository safely. An agent claims a scope, commits in its own worktree and submits.
Porch: add home-gateway Lab integration with duplicate-delivery and forged-sender checks
Porch is a permissioned gateway between AI agents and a connected home. Agents connect to it over MCP, each with its own key.
GridPact: add A2A adapter and Path-mode duplicate and conflict test
GridPact is a peer-to-peer energy-trading protocol for grid agents: agents trade over an MCP server with a locked stake, a hash-chained ledger, deterministic settlement and an auditor that can challenge over-declared capacity.
SharpeShooter: add bounded research workflow and duplicate-delivery test
This adds a focused external-provider integration in which Nanda Town requests a bounded structured research task rather than using a generic quote trigger.
ActionLedger: Track sellers (deterministic and Claude) with ledger-backed external writes under Town's faults
[ActionLedger](https://github.com/Cecilia1220/actionledger) is a small service an agent calls instead of calling a side-effecting tool directly.
Agents Gym: add scorecard-attested trust layer with tampered-card and gym-down tests
[Agents Gym](https://agents-gym.vercel.app) runs a candidate agent through a replica of a real workflow twice: once as is (world A) and once with exactly one controlled change (world B).
FactoryBridge: test real planner fulfillment through HTTP mailbox
FactoryBridge coordinates a simulated hardware order across print, PCB, purchased parts, inspection, assembly and shipping. This PR adds a bounded integration example, adapter, instructions and recorded test artifacts for Nanda Town.
Bazaaro: add registry integration (bazaroo.v1) with name-taken refusal scenario
Bazaaro is a marketplace where AI agents discover, evaluate and hire specialized intelligence (currently medical specialist agents, plus a primary-care agent that assembles a team).
fix(verify): replay bundles judged by a plugin validator
A Lab scenario can bring its own validator through plugin_files (or nandatown run --plugin FILE), but nandatown verify replays the evaluator without loading it.
Agora (formerly Adgentic): run the exchange against Track sellers with a fault-profile compatibility matrix
Adgentic is an exchange for shopping agents. A buyer agent posts its owner's need, seller agents return signed, structured offers, Offer Review checks them against each seller's registered records and strips prompt injection, and the buyer…
Trajano: B2B procurement deal over MCP with lost-accept idempotency and over-mandate checks
[Trajano](https://github.com/augustoal/trajano) is an agent-native B2B procurement marketplace. Buyer and seller agents trade only through its MCP tools, using typed messages (rfq, offer, counter, accept, withdraw, order updates).
AgentTab: add credit-purchase seller integration and charge-once tests
A Track seller (examples/agenttab_seller.py) that buys one price lookup on its AgentTab credit line, over AgentTab's HTTP API, before it answers each quote. examples/agenttab_check.py compares Town's evidence with AgentTab's ledger.
Agent-Hunter: add trial-then-hire and provider invoice-extraction A2A path profiles
[Agent-Hunter](https://github.com/chris-a-ackerman/agent-hunter) is a hiring service for AI agents. Before a requester agent hands off an invoice-extraction job, Agent-Hunter reviews the job, gives every candidate provider agent a short…
FairClear: add sealed second-price procurement integration and forged/late-bid tests
FairClear is a clearing service for AI agents: instead of haggling, a buyer agent opens a sealed request for quotes (RFQ), each seller agent submits one sealed bid, the lowest bid wins and is paid the second-lowest bid (Vickrey RFQ), and…
JANG: delegate multi-term negotiation in a Lab marketplace, with mandate-violation, tampered-agreement, no-ZOPA and service-down checks
[JANG](https://github.com/byujan/JANGmarket) is negotiation infrastructure for autonomous agents. An owner signs a mandate saying which line their agent may not cross on each term.
Remembrane: redirect handshake for scoped health context over the mailbox and A2A, with adversarial-requester and fail-closed tests
Remembrane is a patient-owned health memory for AI agents: other agents get a purpose-scoped view of a person's
TermsMet: add MCP acceptance workflow and lost-reply recovery check
TermsMet is a shared acceptance and conditional-delivery service for buyer and provider agents. The two parties agree on one versioned CSV-to-JSON task.
Leal: add recipe-service integration and sponsorship-bar screening test
Leal helps people trust their agents with jobs that matter. Its first product is recipes: a job someone already figured out, written down so any agent can follow it while it works.
ProbiAI: insured ledger plugin with a double-payment check against Town's ledger
ProbiAI is insurance for AI agents: an agent's own runtime sends ProbiAI a copy of each action, ProbiAI keeps a
AgentSchool: add a credential-verifying trust layer and a revoked-credential test
[AgentSchool](https://github.com/justinspar/agentschool) trains AI agents in one skill, examines them on work they have not seen, and issues a signed credential that other agents check before handing over a job.
SimOps AI: five construction trades coordinate Week 9 through a referee service, with duplicate, late-accept and forged-seat tests
SimOps AI is a simulation environment and coordination layer for construction AI agents. Each trade's agent negotiates one shared schedule, and a neutral referee checks every move against site rules (zones, crane, crew limits, trade order)…
TrueMenu: add verified-quote seller adapter and stale-data compatibility check
[TrueMenu](https://github.com/IntelCodes/truemenu) is a service that ordering agents call before they quote a customer. It returns menu items, prices and open-now status that come from the restaurant, labels each restaurant's data as…
Taste Twin: add shared-taste MCP integration and duplicate-choice test
Taste Twin is a shared taste memory for AI agents. Any agent acting for a person can call it over MCP to rank cafe options the way that person would.
Synapse: verify a consent-scoped record handoff
Synapse gives agents scoped access to a shared record with source citations. This PR adds a NANDA Town workflow that passes a cited synthetic mobility claim between two participants and verifies what arrived against the record.
Open Trading Desk: MCP trading workflow with a lost order confirmation test
Open Trading Desk is an MCP server with three shared services for trading agents: Signal Exchange (provider registry and hash-chained view log), Scorekeeper (provider track records), and Risk Gate (the only path to a simulated broker).
FlexGrid: every node an agent, 8 days against their rules control, with a replayed-id failure case
FlexGrid is a demand-flexibility market for a community of homes and buildings. Seller nodes offer load cuts, a coordinator pools them into one sale, two stand-in buyers order and settle on signed meter readings, and every message is a…
Stellar Compliance AI: add REDECO compliance-agent A2A integration with fault and fail-closed tests
Stellar Compliance AI prepares answers to REDECO consumer complaints (CONDUSEF, Mexico) for a human compliance officer at a non-bank lender.
TrailProof: run an agent purchase and its dispute under Track, cross-verified against a shared ledger
TrailProof is an append-only commitment ledger for purchases made by AI agents, exposed as an MCP server (https://github.com/Andreyivvi/TrailProof).
Add opt-in shared-dependency resilience experiment for Lab
Different registered agents can share the same backend, so switching agent identities may not escape a correlated failure.
fix(dashboard): stop probing submitted links and bound catalog submissions again
fix(dashboard): stop probing submitted links and bound catalog submissions again
Add two-agent web-evidence prompt-injection experiments with optional defenses
This PR adds a local, reproducible two-agent experiment for studying indirect prompt injection through webpage evidence. A read-only Collector extracts visible page content and passes an evidence packet to a separate Answerer, which…
fix(dashboard): expire the likes cache on write so a reload shows the vote
Fixes #186.
Add missed-ack-escalation scenario: 3 dropped prepare_acks force retr…
New scenario testing consensus recovery when acknowledgments are dropped.
examples(llm_market): model-driven buyers in a sealed-data market (stacked on #298)
Depends on #298. Merge after it. This branch is stacked on #298, so until #298 merges the diff also shows its commits. The only new commit is the last one, examples(llm_market). It was moved out of #298 at review, unchanged.
fix(evaluator): do not pass `correct` on an assertion the run contradicts
…icts
Report: attribute the processed claim to the seller's acknowledgement
Town reads the seller’s recorded acknowledgement but doesn’t observe the seller applying the task. This rewords the processed stage so the report says what the evidence is, and adds a test for it.
[Fellowship Challenge] Add terminal Contract-Net task cancellation
Submission for the Hexaware x Project NANDA Fellowship Technical Challenge
[Fellowship Challenge] negotiation: posted-price clocks, Dutch and IOC-limit rising, that hold a delayed accept to a safe price
Submission for the Hexaware x Project NANDA Fellowship Technical Challenge
[Fellowship Challenge] payments: a duplicated award settles once and leaves a verifiable replay record
Submission for the Hexaware x Project NANDA Fellowship Technical Challenge
[Fellowship Challenge] privacy: redact.values.v1, a secret copied under another field name stays out of the evidence
Submission for the Hexaware x Project NANDA Fellowship Technical Challenge
[Fellowship Challenge] Lab: bounded voting finality under delayed ballots
Submission for the Hexaware x Project NANDA Fellowship Technical Challenge
[Fellowship Challenge] payments: add bound.v1, escrow that pays only the seller it names
Submission for the Hexaware x Project NANDA Fellowship Technical Challenge
[Fellowship Challenge] Lab: prevent a losing bidder from being paid for a supply-chain part it was not awarded
Submission for the Hexaware x Project NANDA Fellowship Technical Challenge
[Fellowship Challenge] payments: hashlock.v1, sealed digital goods and payment change hands in one step
Submission for the Hexaware x Project NANDA Fellowship Technical Challenge
[Fellowship Challenge] haggle.expiring.v1 - an offer lapses; a stale acceptance is refused
Submission for the Hexaware x Project NANDA Fellowship Technical Challenge
[Fellowship Challenge] Coordination: one-shot ContractNet finalization under duplicate delivery
Submission for the Hexaware x Project NANDA Fellowship Technical Challenge
[Fellowship Challenge] Expiring memory (ttl.v1) for the memory layer
Submission for the Hexaware x Project NANDA Fellowship Technical Challenge
[Fellowship Challenge] Partition fault: consensus stays safe while the town is split
Submission for the Hexaware x Project NANDA Fellowship Technical Challenge
[Fellowship Challenge] Track: a poison request is redelivered forever; claim + reproduction
Submission for the Hexaware x Project NANDA Fellowship Technical Challenge
[Fellowship Challenge] payments: leased escrow, a stranded hold refunds itself when its lease ends
Submission for the Hexaware x Project NANDA Fellowship Technical Challenge
[Fellowship Challenge] Corrupt fault: prove signed message bodies survive tampering in transit
Submission for the Hexaware x Project NANDA Fellowship Technical Challenge
LOOM: add task-scoped health-context access integration with overreach, expiry and portal-down tests
What this tests: the agent-to-service interaction where a Town agent working for a person asks LOOM for that person's health context.
Add Byzantine fault tolerance scenarios for grid frequency regulation
consensus.yaml only tolerates dropped acks (crash/omission); nandatown has no Byzantine model, and the README explicitly disclaims byzantine_agents as "not modeled." This adds one.
[Fellowship Challenge] Communication : Message Integrity via SHA256 Checksums
Submission for the Hexaware x Project NANDA Fellowship Technical Challenge
feat: add network partition transport plugin and scenario
Adds a partitioned.v1 transport plugin that splits agents into isolated groups. Messages crossing group boundaries are silently blocked and recorded as message_partition_blocked events.
feat: add Protocol Failure Matrix for systematic fault injection across scenarios
Introduces a reusable mechanism for testing protocol-layer failures and recording their downstream effects across (scenario × fault) combinations.
fix(receipt): refuse Town Proof when coverage is incomplete
Town Proof previously required at least one tested stage but did not reject a nonempty coverage.not_tested list. A passing direct A2A run without a pinned AgentCard digest could therefore leave descriptor_consistency untested while still…
feat(agent-test): sign test results and verify them offline
A town.test-result/1 document binds what it ran against: the profile by profile.digest, the run by execution.seed and execution.scenario, the trace by an artifact digest. The document carrying those bindings is itself unsigned.
fix(metrics): don't report zero validators as a passing verdict; guard registry drift
validate_events does VALIDATORS.get(scenario_type, []), and metrics.validate_protocol reported all_passed = all(r.passed for r in vresults).
feat: TrustGuard trust plugin — ELO reputation + risk scoring + denylist + Sybil resistance
Trust layer plugin replacing \score_average\ (running mean) with security-grade agent reputation.
trust(parc): anchor behavioral_merkle_root into a capsule, closing the equivocation gap.
cc @Skyrider3 @dhve @Sharathvc23 — wiring PARC (#63, #66) to the capsule/anchor machinery (#197, #200) that docs/layers/trust.md already named as the next step.
Payments: Prava plugin + policy_commerce scenario
Adds packages/nest-plugin-prava, a Payments-layer plugin that moves real
Hackathon/chainaim sathya prava payments
What it does:
feat: Add Agent Heist scenario with Prava payments plugin
- Agent Heist Scenario: Adversarial marketplace where 8 buyer agents race to purchase Caesar Salad ingredients while 3 of 7 merchants deploy social engineering attacks (prompt injection, price inflation, urgency overcharge)
feat(payments): add nest-payments-prava plugin & CapSpend scenario
This PR introduces the Prava Agentic Payments Adapter (nest-payments-prava) for the NandaHack x Prava Agentic Commerce Hackathon. It enables any Nanda Town scenario to use Prava's Agentic Payments Sandbox via Visa Network Tokens.
feat(payments): Prava agentic-payments adapter + trust gate + adversarial validator
In one line: the first Nanda Town payments plugin that can *refuse a bad payee* — a reusable Prava adapter (plus a validator that proves the trust gate holds) that gives every Nanda builder safe agent-to-agent money, which the default…
feat(payments): Prava sandbox payments adapter [chijian28]
- Adds nest-plugins-prava, a reusable NANDA Town payments plugin that implements quote / pay / verify_payment / refund against Prava's Agentic Payments Sandbox.
soham109: Prava multi-principal group mandates
This contribution adds nanda-town-prava-group, a native NANDA Town payments plugin for multi-principal merchant purchases.
Add Prava Payments Plugin for NANDA Marketplace
Complete payment processing plugin for NANDA ecosystem.
Add reusable Prava payments adapter for NANDA Town
Adds an installable nest.plugins.payments adapter backed by Prava Session APIs. It implements quote, pay, verify and merchant-owned refund handling, plus a one-payment interactive sandbox scenario and tests.
bounded_delegation: cap chain depth, let the revocation set forget
delegatable and mesh_revocable get the crypto right. They just never stop growing — in two directions — and neither growth breaks an invariant, which is why the existing validators wave it all through.
A multi-attribute negotiation extreme strategy
This PR adds a multi-attribute negotiation strategy for price and delivery deadline.
Guard SkillMD submissions against duplicate creation
Fixes the duplicate-submission bug reported in #203 — one Submit click
recommender
No description provided.
feat: implement concrete collective memory layer and protocol tests
Implements a concrete CollectiveMemory layer class and adds corresponding unit tests to verify proper integration, layout structure, and protocol validation.
Fix dashboard: render all panels on file load, and correct >100% success rate
Two independent bug fixes in apps/dashboard/index.html, both affecting any scenario opened via nest dashboard <trace.jsonl> (not specific to any one scenario). No behavior change to scenarios, plugins, or the CLI — this is display-only.
Add `versioned` comms plugin with schema-versioned wire envelopes and forward-compatible unknown-field preservation
This PR adds a new reference comms implementation that ships a schema-versioned wire envelope for Nanda Town. It enables safe rolling upgrades by preserving unknown fields from newer-minor peers and rejecting unknown-major versions instead…
hackathon/daksh-orset-memory: OR-Set CRDT memory plugin — concurrent adds survive
[[02 — Conflict-free shared memory under concurrent writers](https://claude.ai/cowork/docs/hackathon/problems/02-memory-crdt-lww.md)](docs/hackathon/problems/02-memory-crdt-lww.md)
relay: Fix HotStuff locked-QC bypass via omitted justify_qc
If you operate BFT relay infrastructure in production, code review isn't where you find safety bugs — you find them by asking "what does a leader that's lying cost me, and does anything notice?" That's the audit I ran against…
cover gossip registry partition honesty against the real simulator (problem 06)
<html>
Redeem_Grimm: extract reusable heartbeat authenticator, make the forgery discriminator a component swap
This is a direct follow-up to the review note on #152. That review approved the signed-heartbeat forgery defense but flagged two structural points: the authenticated-heartbeat logic lived inside the failure_detection scenario builtin…
docs(registry): capability-conformance plugin note for #159
verified_capabilities plugin, validator, and scenario without a
fix(lint): E501 on main — wrap 103-char test signature from #184
make ci-local currently fails on main itself: ruff check reports E501 (103 > 100) at packages/nest-plugins-reference/tests/test_negotiation_determinism.py:88, introduced in #184.
hackathon/john-kioko-weighted-trust: WeightedTrust plugin (v2 — discriminator + determinism + CI)
Addresses all feedback from the v1 review on #143.
Make DataFacts freshness deterministic
- Add an injectable clock and configurable freshness window to the DataFacts v1 reference plugin.
payments/prepaid_credits: report REFUNDED status and burn refunded refs
Layer 7, Payments — the prepaid_credits reference plugin.
Add streaming per-tick metered payments plugin (payments/streaming)
- Implements open_stream/close_stream per spec 03-payments-streaming-x402
memory: add OR-Set (observed-remove set) CRDT plugin
Adds a state-based OR-Set CvRDT to the memory layer, registered as memory/or_set. Concurrent adds to a key converge to the union (add-wins), unlike lww_register which reduces concurrent writes to a single winner.
fix(nest-plugins-reference): register missing reference plugin entry …
The reference package only exposed a handful of plugins as entry points (memory/lww_register, privacy/hybrid_x25519, datafacts/cid_facts, and the failure detectors).
Fix reputation scoring validator dead loop
Fixes #98
action-state-group: capsule-emit trust layer — verifiable, anchored records for NANDA agents
Primary submission — supersedes #54 (same commits, branch renamed to hackathon/stevenmih-capsule-trust per review process requirement). Review history is in #54. Companion PR: #55 (Problem-09 SD-JWT privacy plugin, secondary submission).
rahulsharma-sys: sealedbox — nonce-misuse-resistant hybrid privacy (Problem 09)
Reworked sealedbox privacy plugin (Problem 09). This supersedes #36, which
build: migrate deprecated tool.uv.dev-dependencies to dependency-groups.dev
uv prints a deprecation warning on every command: tool.uv.dev-dependencies
andrea-cola -> ed25519_recoverable: time-locked rotation + K-of-N social recovery (Layer 3 identity)
* Layer: identity (Layer 3)
trust-security-engineer: attested_peering operator delegation is not key-bound — trusted-operator impersonation
The attested-peering trust plugin gates reputation on a "who do you work for?"
Problem 10: Add namespaced BFT quorum protocol with signed equivocation exclusion and view change
This PR adds a new, submission-owned BFT quorum consensus implementation for Problem 10.
feat(registry): add dp_bloom, a differentially private registry with a membership-inference validator
This PR adds dp_bloom, a differentially private registry. It serves discovery exactly as the in-memory reference does and publishes a separate membership index that satisfies ε-differential privacy.
hackathon: embx_semantic memory plugin (Layer 10) -- recall by meaning
A new embx_semantic plugin for Layer 10 (Memory). The default blackboard is an 80-line shared dict -- read a value under a key, get it back under the same key, nothing else.
tony-nexartis: delegated_admission — evidence admission gated on human-anchored, cascade-revocable delegation grants (production port)
A trust-layer plugin, delegated_admission, that gates reputation
fix(identity): ed25519_rotating replay-corruption + destructive re-registration hardening
Red-team remediation of the merged ed25519_rotating identity plugin (#25), following the precedent of #108 red-teaming merged work.
feat: implement delegatable capability tokens with cascading revocation.
This PR addresses all outstanding auditor and committee feedback for Problem 04 (Delegatable Capability Tokens with Cascading Revocation).
rajpatilrobotics: fail closed on empty scenario traces
NANDA Town is a protocol test rig, so a successful validation must be backed by actual trace evidence.
auth-safety-engineer: DelegatableAuth with scoped delegation, cascading revocation, and deterministic validators
Persona: auth-safety-engineer
auth-safety-engineer: DelegatableAuth with scoped delegation, cascading revocation, and deterministic validators
Persona: auth-safety-engineer
Maintenance/fix reputation validator
Summary
adarx: tiered-discount negotiation plugin with strategic bargaining
A third plugin for the Negotiation layer — adarx — alongside the existing alternating_offers (Rubinstein patience-discount) and pareto (utility-based trade-off) reference implementations.
hackathon/mwarsss-delegatable-auth
Implements Problem 04 → Auth Capability Delegation from the NandaHack charter. Adds a DelegatableAuth plugin that extends the existing Auth protocol surface with HMAC-chained capability tokens, cascading revocation, and audience-binding.
stripe-engineer: streaming payments — idempotency, conservation invariants, and a working end-to-end scenario
Supersedes #116 (renamed to hackathon/stanleyoz-streaming-payments per the charter's hackathon/<handle>-<theme> branch-naming rule; same author, same content plus the fixes requested in that review thread).
distributed-sys-engg: deterministic coordination round ids (ADR-004 fix)
Persona: distributed-sys-engg — reliability/determinism focus.
docs(trust): document attested_peering; drop dead attested_peers accessor
Follow-up polish to #149 (merged, thank you!). Two small, non-blocking items that
capability-systems-engineer: delegatable capability tokens with cascading revocation
Solves docs/hackathon/problems/04-auth-capability-delegation.md.
: add replay-safe comms plugin with sequence-bound envelopes
A new replay_safe comms plugin that extends authenticated with replay attack detection via per-peer monotonic sequence numbers. This PR closes the gap explicitly called out in PR #114:
Nandatown: composable middleware, distributed HTTP hardening, skills API security, and local deploy tooling
| Tier | Files (review first) |
Bind score-average trust evidence to subjects
ScoreAverageTrust trusted two caller-controlled references to the subject independently. A malicious caller could report evidence about victim while mutating innocent, and an attestation could pair an arbitrary agent argument with a claim…
hackathon/john-kioko-weighted-trust: WeightedTrust plugin
A new trust plugin that improves on ScoreAverageTrust with:
auth-engineer: delegatable capability tokens with cascading revocation
Solves hackathon problem 04 — Delegatable capability tokens with cascading revocation.
auth-safety-engineer: DelegatableAuth with scoped delegation, cascading revocation, and deterministic validators
Persona: auth-safety-engineer
Nandatown: composable middleware, distributed HTTP hardening, skills API security, and local deploy tooling
| Tier | Files (review first) |
feat(trust): implement 2-pass SybilResistantTrust plugin
This PR implements a robust, security-hardened 2-Pass Eigen-weighted Trust & Reputation plugin (sybil_resistant) under the Trust layer, along with a dedicated test scenario and active decision-gating to prevent Sybil attacks, collusion…
feat(memory): implement OR-Set CvRDT memory plugin with tick clock sy…
This PR implements a robust, API-compliant state-based OR-Set (Observed-Remove Set) memory plugin to resolve Problem 02 (memory-crdt-lww).
settlement-engineer: split_settlement — weighted multi-payee fan-out settlement with conservation-proof validators
The merged payments plugins each move value in one shape:
Improve marketplace PR short descriptions
- Skip Markdown bullet lists when deriving short descriptions for hackathon submissions.
feat: TownOps triage scenario with multi-agent coordination
Adds a complete town operations triage scenario for the NANDA Town hackathon.
feat: add structured receipt field to Evidence type
feat: add structured receipt field to Evidence type
auth: macaroon-style delegatable capability tokens with cascading revocation
The default auth plugin (nest_plugins_reference/auth/jwt_auth.py) is 106 lines of HMAC-signed token issuance. It has no notion of delegation: JwtAuth._revoked: set[str] tracks revocation by exact token string, with no parent-child…
audit-engineer: intent-gated datafacts — pre-publication intent gate (problem 08)
Layer: datafacts · Problem: 08 · Persona: audit-engineer · Author: Midhun Raj Charles (midhunrajcharles) — same name as the Phase 2 skills submission ([AgentIntent in the registry](https://nandatown.projectnanda.org/skills), entry…
SkillMD API reachability validation
This Phase 1 warm-up improves the SkillMD submission API so programmatic submissions get the same basic safety and reachability behavior as the form path.
fix: enforce reputation score property in validate_reputation_scoring
fix: enforce reputation score property in validate_reputation_scoring
Harden LWW register state validation
Harden the memory layer's LWW-register CRDT merge boundary against malformed or adversarial replica state.
Manifest-bound delegatable auth
This submission targets Problem 04, Delegatable capability tokens with
Remove dead code from reputation scoring validator
This PR removes the unused dead loop in validate_reputation_scoring and updates the related comments to reflect the validator's current behavior.
maharmuavia (distributed-systems-engineer): or_map -- add-wins observed-remove map CRDT (memory)
Persona: distributed-systems engineer. The thing I care about is *causal*
feat: Add AgentCourt escrow, dispute resolution, and persistent memory services
This PR implements the AgentCourt Escrow & Dispute Resolution service and the Agent Persistent Memory API for Nanda Town.
hackathon/implement NANDA Simulation Sandbox API and serve SKILL.md phase 2
Implementation of NANDA Simulation Sandbox (SaaS). This service allows remote or lightweight AI agents (e.g., OpenClaw agents) to run full NANDA Town simulations, calculate performance metrics, and retrieve adversarial validator results…
Fix: Address consensus feedback, upgrade to BFT, and remove non-deterministic engine
This PR addresses all four critical feedback items from the previous review regarding the NandaQuorum consensus implementation.
LasaKaru : delegatable capability tokens with cascading revocation
Add a macaroon-style auth plugin that lets an agent delegate a strict subset of its capabilities to another agent without going back to the issuer, and that revokes an entire delegation subtree transitively when any ancestor token is…
feat(datafacts): add MerkleDataFacts plugin with Git-style DAG, monotonic hash chains, bloom ancestry, and aliases
Implements Problem 08:
auth-engineer: delegatable capability tokens with cascading revocation
The default auth plugin (jwt_auth.py) issues flat HMAC tokens and revokes by
hackathon/yashjadhav1595-projects : Ed25519 rotating identity with key rotation
The default did_key plugin produces a static, deterministic identity that cannot be rotated — a compromised key remains valid forever.
stripe-engineer: Production-grade streaming payments with idempotency, conservation invariants, and adversarial property tests
This PR hardens the streaming payments plugin against the failure modes that naive debit-credit ledgers silently accept. Every stream mutation is now idempotency-keyed, and the plugin enforces three invariants at the balance level that…
Delegatable capability tokens with cascading revocation (auth)
Solves hackathon problem [04-auth-capability-delegation](docs/hackathon/problems/04-auth-capability-delegation.md) (difficulty: easy).
jmuslu: PN-Counter memory with basis-fusion invariants
This PR adds a memory:pn_counter reference plugin, a memory:basis_gated
Expand behavioral test coverage for the score_average trust plugin
Adds additional behavioral tests for the reference score_average trust plugin.
EFS Scribe delegated auth with offline write receipts
This PR adds auth: delegatable, an HMAC-chained capability-token plugin for Problem 04.
Add policy decision layer for agent actions
NANDA Town has layers for identity, auth, privacy, payments, trust, memory, coordination, and failure detection, but it does not yet have a small deterministic policy decision point for agent actions.
hackathon/tanisha: implemented a custom Rate Limiter layer and reference plugins
Implementation of a custom Rate Limiting Layer (introducing the rate_limit layer, TokenBucketRateLimiter, and LeakyBucketRateLimiter reference plugins) to enforce traffic and request boundaries per agent.
distsys-reliability: or_set memory CRDT + honest-write-liveness red-team of merged lww_register
Persona: distsys-reliability — a distributed-systems reliability engineer.
hackathon/robert-samuel-staking-primitive
What I did
hackathon/dzaky-delegatable-auth: Problem 04 — Delegatable Capability Tokens
Agents in NANDA Town currently have no way to mint bounded sub-credentials and pass them downstream without routing every authorisation request back through the original issuer.
hackathon/Implement delegatable capability tokens and scenario doctor
This PR contains the implementation of Delegatable Capability Tokens with Cascading Revocation (Phase 1) and the Scenario Doctor FastAPI validation service (Phase 2).
hackathon/mwarsss-delegatable-auth
Implements Problem 04 -> Auth Capability Delegation from the NandaHack charter. Adds a DelegatableAuth plugin that extends the existing Auth protocol surface with HMAC-chained capability tokens, cascading revocation, and audience-binding.
hackathon/yashjadhav1595-projects: Macaroon-style capability tokens with cascading revocation (Problem 04)
The default jwt_auth plugin handles root token issuance, but lacks any mechanism to safely delegate scoped access down a hierarchy, and critically, it lacks parent-child revocation.
Fix: Address consensus feedback, upgrade to BFT, and remove non-deterministic engine
This PR addresses all four critical feedback items from the previous review regarding the NandaQuorum consensus implementation.
Add delegatable capability tokens with cascading revocation (auth layer)
The default jwt plugin can't delegate — there's no way for an agent
Hackathon/amit auth delegation
DelegatableAuth Plugin Problem 04
metrics: fix delivery_rate/success_rate exceeding 1.0 on broadcast-heavy scenarios
Fixes delivery_rate (and its success_rate alias) returning a value greater than 1.0 on any scenario that uses broadcasts. A delivery "rate" above 1.0 is impossible and misreports every broadcast-heavy run.
memory: add mv_register, a multi-value register CRDT that never drops a concurrent write
Adds mv_register, a Multi-Value Register CvRDT for the memory layer. The existing lww_register is last-writer-wins: on two concurrent writes to a key it keeps one and silently drops the other.
hackathon/superheroin888: delivery-gated streaming payments with mid-stream cancellation (problem 03)
Problem 03 — streaming pay-per-second payments with mid-stream cancellation (payments layer). One problem, one layer.
feat(auth): Macaroon-based delegatable tokens with cascading revocation
This PR solves Problem #04 by adding the macaroons auth plugin.
theodornengoy (security-engineer): strict holder-derived capability delegation
Solves problem 04, auth capability delegation, as a distinct implementation
registry: DNS-style caching resolver with TTL, negative caching, and self-evicting agents
Adds resolver, a DNS-style caching resolver registry plugin: per-record TTL, negative-result caching, and self-eviction of stale/dead agents.
macaroon-delegation: delegatable capability tokens with cascading revocation
The default jwt auth plugin tracks revocation by exact token string and has no parent/child relationship, so it cannot model the common pattern: A holds a root token, A mints a narrow short-lived sub-token for B offline, and revoking A…
[NandaHack] privacy-comms: Tor-style Onion Routing anonymity network
🧅 Layer 02 (Communication): Tor-style Onion Routing Network
feat: implement Problem 04 Delegatable Auth with simulation and valid…
feat: implement Problem 04 Delegatable Auth with simulation and valid…
reputation_negotiation scenario — trust-gated negotiation…
… with reputation threshold
nandha-datafacts-verifier: Delegatable capability tokens …
Problem #4: Delegatable Auth (True Macaroon-style Capability Tokens)
privacy: Trust-Gated Selective Disclosure
Privacy (Layer 11) — Reference plugin implementation of Trust-Gated Selective Disclosure.
/tani-decaying-memory
I have implemented a new plugin for Layer 10 (Memory) called DecayingMemory. This introduces a Time-To-Live (TTL) mechanism to agent memory, allowing for "biological-inspired forgetting" within the Nanda Town simulator.
fix: reputation scenario fails its own validator + dashboard SSRF
Running nest run reputation with the default seed of 42 produced a trace that failed the reputation_scoring validator. The committed demo trace had the same issue.
cloud-post-code: streaming payments scenario factory (#03)
Problem: 03-payments-streaming-x402 — Streaming pay-per-second payments with mid-stream cancellation
payments-risk-eng: BudgetLimitedPayments — cross-source spend-cap guard
Persona: payments-risk engineer.
Hackathon/a2r commerce checkout frontier
Agentic checkout is multi-attribute. A shopping agent negotiating a purchase is trading price against delivery deadline ("I'll pay a premium for expedited"), under a constraint no utility weight can express: a hard wallet budget.
geick: typed CRDT memory for concurrent agent writes
This submission adds a typed_crdt memory plugin for problem 02, concurrent memory under unreliable delivery.
feat(registry): persistent Cloud SQL and Redis registry backends
The default in_memory registry is a shared dictionary — single-process, non-persistent, and transparent to the simulator's partition injection.
andrea-cola: medical data exchange scenario for hybrid_x25519 privacy
- Adds a new Tier-1 scenario, medical_data_exchange.yaml, exercising the hybrid_x25519 plugin's selective-disclosure path in a medical-data workflow (1 hospital + 3 insurers + 1 auditor).
capsec-engineer: offline-attenuating capability tokens with epoch-fenced cascading revocation
The reference jwt auth plugin signs exact tokens and revokes exact token strings. That is useful scaffolding, but it is not a capability system: a holder cannot attenuate authority offline, and revoking a parent does not invalidate…
maurathat: sic_facts — serialization-invariant content addressing (closes the provenance-laundering gap in byte-level CIDs)
TL;DR: cid_facts prevents name substitution, but not identity forking under semantically equivalent serialization. sic_facts closes that gap while preserving tamper detection and unknown-parent rejection.
chainaim-agentID-engineer: Ed25519 identity with KERI pre-rotation and hijack-resistant key rotation
Problem #05 — identity / Ed25519 key rotation. Adds a KERI-style pre-rotation identity plugin at ("identity","ed25519_prerotation"), building on the merged reactive-rotation plugin (PR #25) to defend a third attack class — rotation hijack…
fayner-delegatable-auth: macaroon capability tokens with cascading revocation (problem 04)
Problem: #04 — Delegatable capability tokens with cascading revocation (auth layer).
brettleehari: fair-ordering a red-team of the coordination layer + the primitive that closes it
[Hackathon] brettleehari: fair-ordering — a red-team of the coordination layer + the primitive that closes it
Hackathon/disaster response engineer KumbhNet 2027
Handle: disaster-response-engineer
vladimirputkov: verified_registry — signature-verified registration with typed rejections and a Sybil scenario
Branch: hackathon/vladimirputkov-registry-integrity
RoboAgent Guard for Safety and Privacy Validation
RoboAgent Guard targets NandaHack Phase 1 Problem 09: privacy.
cloud-post-code-pareto-negotiation: multi-attribute bargaining, Pareto frontier, agents trading off price vs deadline
The default alternating_offers plugin bargains over price only. This PR adds a second plugin, pareto, where agents bargain simultaneously over price, deadline, and quality and converge toward Pareto-optimal agreements: deals where neither…
chainaim-nisha : Outcome-Verified Settlement: per-tick payments gated by delivery, integrity, and conformance truly allowing for outcome versus tokens
Problem: [03 — Streaming pay-per-second payments with mid-stream cancellation](docs/hackathon/problems/03-payments-streaming-x402.md)
hiten: add configurable provenance_supply_chain_multi scenario
This is a focused follow-up to my previous hackathon PR (#39). Following the merge of the DataFacts implementation in #31, this PR contributes the independent scenario enhancement suggested during review by introducing a configurable…
interview-integrity: Problem 08 (datafacts) — ogha_facts plugin: audience ACLs + pre-hash PII redaction + screening-evaluation scenario & validators
I run the AI services for a verified-screening business — the filter at the *front* of a client's hiring funnel. A human interviewer conducts a recorded screening interview; the audio is transcribed, and the interviewer files an evaluation…
srcJin: Problem #10 — ResonanceBFT, pentadic partition-tolerant BFT consensus for social agent networks
Layer: coordination
distributed-systems-architect: private_commerce — four-layer cross-plugin composition with joint validators
None of the ten listed problems — this PR targets the gap they leave behind. Every merged hackathon plugin (gossip registry #24, streaming payments #21, hybrid_x25519 privacy #28, agent_receipts trust #26) was designed, tested, and…
theritwik: delegatable capability tokens with cascading revocation
Implements hackathon problem #04 (docs/hackathon/problems/04-auth-capability-delegation.md):
Problem-09 privacy: SD-JWT capsule selective-disclosure plugin
Implements the SD-JWT salted-hash commitment model from
action-state-group: capsule-emit trust layer — verifiable, anchored records for NANDA agents
capsule-emit-nanda — a drop-in NANDA trust plugin backed by the [Agent Action Capsule](https://github.com/action-state-group/agent-action-capsule) spec (IETF draft).
varuni7: delegatable capability tokens with cascading, partition-tolerant revocation
Problem 4 (auth). The default jwt plugin cannot delegate a subset of a
scenarios: provenance_supply_chain_linear — deep linear provenance chain (follow-up to #35)
Follow-up #2 to #35 (closed as a duplicate of #31), the distinct-topology scenario @dhve flagged as worth landing independently.
datafacts: first-class typed parents field on DatasetMetadata (follow-up to #35)
Follow-up #1 to #35 (closed as a duplicate of #31), the typed-lineage-field piece @dhve flagged as a clean, independent improvement.
hiten: add DidKeyIdentity helper accessors
This is a focused follow-up to my previous hackathon PR (#39). Following the merge of the DataFacts implementation in #31, this PR contributes an independent identity-layer API improvement by adding two small helper accessors to…
hiten: add typed FreshnessProof model
This is a focused follow-up to my previous hackathon PR (#39), implementing the reviewer-suggested FreshnessProof API improvement after the DataFacts work from #31 was merged.
chainaim-sathya: feasible-frontier audit + generic-N multi-attribute negotiation with a heuristic/Bayesian opponent-model seam (price + deadline + quantity-3 axis active)
AFFIRMS DEAL ( Agreed-Fair-Distance) - CONFORMS DELIVERY - EFFECTS SETTLEMENT IN A
PBFT coordination: BFT consensus, view-change, adversaria…
Problem 10 — Partition-tolerant BFT coordination (PBFT)
NandaQuorum: implement NandaQuorum consensus protocol and scenario testing
In decentralized environments like Nanda Town, nodes must achieve consensus on state transitions and proposals in a deterministic, safe, and live manner despite potential node crashes or delayed network packets.
The_Dealmakers: multi-attribute bargaining, Pareto frontier, agents learning to trade off price vs quantity
The default alternating_offers plugin bargains over price only — it reads one number from Terms and ignores everything else.
The Dealmakers: multi-attribute bargaining, Pareto frontier, agents learning to trade off price vs quantity
Multi-attribute negotiation scenario where buyer-seller pairs bargain simultaneously over price × quantity. Unlike price-only strategies, agents must explore both axes to reach the Pareto frontier — a joint optimum invisible to…
platform: audit remediation (trace schema, CI, ADRs, telemetry)
- Enterprise audit doc + simulation plugin UserWarnings (jwt, did_key, noop)
hiten: Layer 12 DataFacts — content-addressed provenance …
The existing DataFacts layer could store and retrieve datasets, but it could not reliably answer a critical trust question:
Auth: manifest-bound delegatable capability tokens
Problem #04 — Delegatable capability tokens with cascading revocation.
Add sealedbox privacy plugin (real authenticated encryption)
The default noop privacy plugin is a passthrough — encrypt/decrypt return the input unchanged and proofs always verify. This PR adds sealedbox, a real implementation of the Privacy layer: multi-recipient sealed-box encryption with a fresh…
provenance-engineer: content-addressed datafacts with provenance chains & signed freshness proofs
Problem 08 (datafacts). The default datafacts_v1 plugin addresses datasets by publisher-chosen name (df://<name>), grants access unconditionally, and "proves" freshness with a bare time.time() check.
feat(datafacts): implement content-addressed plugin
This PR implements Problem 08: Content-Addressed Datasets with Provenance Chains and Freshness Proofs for the NANDA Town hackathon.
feat(examples): Agent Action Capsule trust layer + Stripe payment demo
Lands a working capsule-emit-nanda plugin package so NANDA users can
feat(auth): implement (Problem 4) delegatable capability tokens with cascading revocation
I implemented Macaroon-style delegatable capability tokens for the auth layer (Problem 4).
TrustGuard trust plugin — ELO reputation + risk scoring + denylist + Sybil resistance
TrustGuard trust layer plugin replacing score_average (running mean) with security-grade agent reputation.
Problem 02: LWW-Register CRDT memory plugin for concurrent writers
The default blackboard memory uses last-writer-by-arrival with no causal history, so replicas that observe the same writes in different orders silently diverge.
NandaHack Warm-up: Streaming Payments Plugin
Streaming payments plugin for NANDA Town. 5 buyers open per-tick payment channels to 5 sellers with rolling streams and 5% message drop, matching Problem #03 exactly.
feat(comms): versioned message envelopes with forward/backward compat…
- plugin: VersionedComms (registered as comms/versioned)
google-staff: netem transport + tail-latency metrics + SLO validators
Layer 1 — Transport. A netem-style alternative transport plugin, plus the observability and assertion machinery a real SRE would need to use it: tail-latency metrics and per-scenario SLO validators.
meta-backend: realistic transport plugin (latency, jitter, queueing, loss)
The README is candid about it:
cybersec-blackhat: dpop_jwt auth plugin + security validators
cybersec-blackhat — senior security researcher. I look at every API and ask "how do I break this?" first, then "how do I harden it?" Auth, identity, trust layers are home turf.
linux-kernel: per-hop latency models for the in-memory transport
Transport (layer 1) — reference plugin extension + simulator wiring.
coinbase-crypto: htlc_escrow payments plugin (hash- & time-locked conditional payments)
The default prepaid_credits plugin transfers funds the instant pay() is
stanford-ml-phd: EigenTrust plugin for the trust layer
Layer 6 (Trust) — adds a second built-in plugin, eigentrust, alongside the existing score_average.
cmu-robotics: sealed-bid auction coordination plugin (first-price + Vickrey)
Layer 8 — Coordination. Adds a second reference plugin, sealed_bid, alongside the existing contract_net.
openai-llm: semantic memory plugin (recall + TTL + LRU)
Layer 10: Memory. The default blackboard plugin is a shared dict — perfect for state-machine agents that already know the key they want, but the wrong shape for the thing LLM agents actually do: *"recall the most relevant past interaction…
mit-undergrad: EigenTrust plugin for the trust layer
Layer 6 — Trust. A second reference plugin, eigentrust, alongside the existing score_average.
harvard-phd: EigenTrust plugin with checkable invariants
Trust layer. Adds eigentrust as a bundled alternative to the
chore: ignore .claude/ harness directory
- Add .claude/ to .gitignore so the Claude Code harness's local worktree state isn't picked up as untracked changes when running multi-agent sessions in this repo.