Code review · CorrectnessThe 200k compute limit is hardcoded and only asserted to be within the 300k ceiling; nothing measures the actual forwarder + PDA derivation + account creation + transfer cost, so an under-estimate would surface as an opaque on-chain compute-exhaustion failure rather than a caught regression.cre-runner/cre/test-workflow/solana.go:33
Code review · TestsThe added assertions only check that the constant is plumbed through to the captured config; they never validate the limit is sufficient or that the mock captures ComputeConfig (no mock/capture change is included in the diff, so this may not even compile).cre-runner/cre/test-workflow/solana_test.go:136-137
Code review · CorrectnessThe PR is labelled as fixing issue #4, which is a placeholder 'test issue', so the reported bug (capability rejecting a nil compute config) is neither described nor reproduced anywhere in the change.
Issue fit · Issue relevanceThe linked issue #4 is a placeholder test issue with no requirements or acceptance criteria, so the change cannot be verified as resolving it.
Issue fit · ValueThe new compute limit is hardcoded to 200k rather than derived from or validated against the referenced cre/simulation-limits.json gas limit, creating drift risk.cre-runner/cre/test-workflow/solana.go:30-33
Issue fit · ValueThe test only confirms the mocked capability receives the compute config; it does not exercise the claimed real capability rejection of a nil compute config.cre-runner/cre/test-workflow/solana_test.go:136-137