A playable city-management sandbox where humans and agents share one live city. WebMCP exposes traffic, budgets, and utilities so every agent action is visible, validated, and reversible.
Aggregate36
Leverage10
Execution9
Impact8
Creativity9
Each criterion 1–10, equally weighted; aggregate is their sum. Ranking is the pipeline's consolidated output.
Origin, access model, eligibility, and substitution are reviewer diagnostics, not judging criteria. Authentication requirements are not penalized.
02 The two blind reviews
Two independent reviewers scored this project blind, from a sanitized evidence packet.
Scores are shown separately so the reasoning stays inspectable. A withheld score means the reviewers differed by more than two points.
Reviewer A (round 1)
confidence 82%
Leverage
9 /10
Structured semantic reads, previews, legal mutations, validation, and reversibility are the product interface for the agent; pixel-driving would be substantially less reliable and could not offer the same shared semantic contract.
Evidence cited
ABOUT calls WebMCP the product interface rather than an automation wrapper.
Fourteen domain tools expose city, metric, area, traffic, budget, mutations, simulation, undo, and related operations.
Agent preview and constraint verification operate on the same deterministic world the human sees.
Execution
8 /10
The packet presents a well-bounded simulator with deterministic rules, broad state tracking, and a coherent agent workflow. No video limits direct observation but the evidence is detailed and internally consistent.
Evidence cited
Demo alive, public repo, and three frame sheets plus packaging screenshot submitted.
ABOUT specifies 24x18 grid, local save/load, metrics, challenge constraints, and reversible actions.
Claims visible improvement requires fresh inspection and 5/5 constraints satisfied.
Impact
7 /10
As a game it offers credible play and learning value, and as a proving ground it makes semantic agent collaboration and tradeoffs concrete; audience value is clear even if niche.
Evidence cited
Playable city-management loop with budgets, traffic, utilities, and challenge constraints.
Shared-world design makes agent actions visible and understandable to players.
Creativity
8 /10
The semantic city as a playable proving ground for agent-native interfaces is a strong synthesis, especially with previews, validation, and reversible live state.
Evidence cited
Treats structured meaning, not pixels, as the agent-facing game surface.
Fresh inspection requirement and constraint verification turn agent actions into a measurable collaboration loop.
Reviewer B (round 2)
confidence 86%
Leverage
9 /10
Live semantic state, validation, previews, deterministic simulation, and reversible mutations make the agent a first-class participant rather than a UI automator.
Evidence cited
About text states WebMCP is the product interface and lists 14 domain tools.
Frames show semantic inspection, congestion visualization, and a previewed infrastructure change.
Execution
8 /10
The city simulator appears coherent and deterministic with visible state transitions, but the frame sheets show inspection/preview more clearly than a completed intervention.
Evidence cited
About text specifies 24x18 grid, metrics, save/load, undo, and challenge constraints.
Three sheets show stable city state, traffic coloring, semantic panels, and preview workflow.
Impact
8 /10
City-management players and agent-interface researchers get a concrete, understandable proving ground for shared-world decisions and constraint satisfaction.
Evidence cited
Frames visibly expose traffic/road and semantic state.
About text reports a goal with 5/5 constraints satisfied.
Creativity
8 /10
A playable deterministic city is a strong and memorable testbed for semantic shared-world agent interaction.
Evidence cited
Human-visible city and structured agent world are explicitly unified.
Preview, validation, reversible operations, and constraint checks are pursued together.
03 Review highlights
Standouts across reviewers
Exceptional alignment between domain semantics, shared state, and WebMCP actions.
Reversible, validated city operations make the agent's contribution legible.
Shared revisioned world with reversible agent actions.
Deterministic, constraint-aware simulation.
Red flags
No submitted video; image inspection was blocked by rate limiting.
Frames show preview and inspection more definitively than completed mutation.
04 Evidence
What each artifact proves is labeled on the artifact itself. A Devpost page capture is packaging evidence, not proof the product runs;
video frames are evidence from the submitted demo, not live verification.
Contact sheets from the ?s video the team submitted. This is what reviewers were shown;
it demonstrates the product in motion but is not independent verification.
· watch the original
EX-V1 Sheet 1 of 3 · reported video evidenceEX-V2 Sheet 2 of 3 · reported video evidenceEX-V3 Sheet 3 of 3 · reported video evidence
