Skip to content

Hackathon #4

Last updated: 2026-06-28

Hackathon #4 centers on agent applications that can delegate work, execute through Cortensor, verify results, and continue workflows. It is the closest hackathon track to the current Corgent, Bardiel, router, and agent ecosystem direction.

Program Summary

Field Detail
Theme Agent Apps: Delegate, Execute, Verify.
Status Planned around Q1 2026; use the latest community announcement for active dates.
Main output Usable agent workflows, templates, validation surfaces, and public-good tools.
Optional surfaces ERC-8004 artifacts, x402, MCP patterns, COR Prover-style validation, and router /validate work.

Workflow Target

sequenceDiagram
  participant User
  participant Agent
  participant Router as Cortensor Router
  participant Miners as Miner Sessions
  participant Validators
  participant Evidence

  User->>Agent: Goal or task
  Agent->>Router: Delegate work
  Router->>Miners: Execute inference
  Miners-->>Router: Candidate outputs
  Router->>Validators: Validate or score
  Validators-->>Router: Verdicts and reasons
  Router-->>Agent: Result plus evidence
  Agent->>Evidence: Store logs/artifacts
  Agent-->>User: Answer, action, or next step

Primary Focus Areas

Track What qualifies
Agentic applications Agents that monitor, act, coordinate, transact, or report with clear tool interfaces.
Delegation and execution loops A full path where the agent plans, delegates to Cortensor, receives results, and continues.
Trust and verification Cross-run checks, rubric scoring, evidence bundles, audit logs, and "why trust this output" surfaces.
Agent tooling and templates SDK wrappers, CLIs, session helpers, structured logging, replay, and evaluation harnesses.
Operational agents Router health checks, latency alerts, incident summaries, support triage, and analytics.
Public-good agents Open bots, limited free tiers, public endpoints, educational tools, and community utilities.

Prize Pool

Placement Prize
1st $1000 equivalent
2nd $800 equivalent
3rd $500 equivalent
4th-5th $300 equivalent each
6th-7th $150 equivalent each
8th-10th $100 equivalent each
11th-20th $50 equivalent each

Prize placement is based on quality, impact, technical rigor, usability, documentation, compliance, and alignment with Cortensor's roadmap. Cortensor may withhold, split, or reallocate prizes when entries do not meet the bar.

Project Ideas

Family Examples
Real operators Release manager, incident commander, QA/regression, community support agent.
On-chain and crypto-native agents Receipt verifier, treasury watcher, governance analyst.
Multi-agent systems Coordinator agent, disagreement resolver, arbitration bundle generator.
Memory and knowledge Project memory agent, knowledge-base builder with citations and confidence.
Tooling and DX Agent starter templates, evaluation harness, observability dashboards.

Deliverables

  • Public repo with permissive license.
  • README with quickstart and runbook.
  • Architecture diagram.
  • Tool list and constraints.
  • Demo link or recorded video with reproduction steps.
  • Runtime proof: transcripts, logs, structured outputs, replay scripts, or test commands.
  • Verification proof: rubric prompts, cross-run checks, validator usage, or evidence bundles.

Evaluation

Criterion Weight
Agent capability and workflow completeness 30%
Cortensor integration 25%
Reliability and safety guardrails 20%
Usability and demo quality 15%
Public-good impact 10%

Bonus consideration can include ERC-8004 artifacts, x402 flows, MCP patterns, /validate usage, on-chain records, tests, benchmarks, and observability dashboards.