The Compiler is the Judge
Read compilation, tests, analysis, deployment and publication as separate results.
Last verified against production: . See current limits and availability.
The verdict comes from the program that ran
The agent writes code and explains its choices. Foundry compiles it, the test runner executes its tests and Slither analyzes the contracts. The pipeline stores these results and uses them to decide whether deployment may proceed. A model saying the work is finished cannot substitute for a missing report.
Repairs have bounded stages. A compiler error, failed test or blocking security finding can give the agent another pass with the actual error output. Required checks run again against the repaired source before a contract can deploy. The Build manual shows the full order and the per-stage evidence.
Read the scope of each result
| Result | What it establishes |
|---|---|
| Compiled | The compiler accepted that source with the recorded settings. |
| Tests passed | The executable tests ran and passed. Their assertions define the behavior exercised. |
| Slither result | Static analysis ran and returned the recorded findings. It is a separate check from tests. |
| Deployment receipt | The recorded transaction created the selected contract on the named chain. |
| Source verification | The explorer accepted the verification attempt when a verified result is recorded. |
| Published | The front-end build was published and its served page returned HTTP 200. |
Check the app's own actions
Open the app and test its required flows on the stated network. A raffle needs ticket purchase, draw conditions and pot claims exercised with the appropriate wallets. A successful compile or hosted page alone cannot establish those interactions.
Keep the exact toolchain result when describing the app to another person or agent. Source verification is attempted after deployment and can fail independently while the deployed contract remains. The console records that distinction.