Formal Verification Tool Output Comparator

WHAT IT DOES

This tool compares outputs from multiple formal verification tools to identify discrepancies, consensus, and verification gaps in smart contract proofs. By aggregating results from Certora, Foundry, Echidna, Mythril, Slither, and other verification tools, it provides a unified view of which properties have been verified, which tools agree or disagree, and where verification coverage is incomplete. The comparator analyzes property-level results—including pass/fail status, violation counts, and execution times—and computes consensus metrics to highlight discrepancies between tools. Results are presented as a clear comparison dashboard showing each property's status across tools, consensus classification (pass, fail, discrepancy, partial-pass, mixed), and detailed breakdowns. This is essential for security auditors validating verification results, developers ensuring comprehensive property coverage, and teams making informed decisions about contract correctness.

HOW TO USE

Connect your wallet (read-only, no transaction required) and select the target network. Enter the contract name and address, then paste the JSON-formatted outputs from your formal verification tools into the text area. The input should be structured as a JSON object with tool names as keys and their results as values. Optionally specify a property filter to focus on specific invariants. Click "compare verification results" and the tool will parse the inputs, extract property-level results, and generate a comprehensive comparison report. Results display a summary with pass/fail/discrepancy counts, a series of result cards showing each property's status across tools, a detailed table of all properties and tool results, and consensus classification for each property.

TECHNICAL MECHANISM

The comparator uses a structured parsing engine that handles output formats from multiple formal verification tools. It accepts JSON input with tool-specific schemas—Certora outputs (properties array with status), Foundry test results (tests array with pass/fail), Echidna property reports (properties with violations), and generic tool outputs (status with metrics). The parser extracts property-level results from each tool, normalizing status values to 'pass', 'fail', 'warn', or 'unknown'. For each property, the comparator calculates consensus by analyzing the distribution of statuses across tools: unanimous pass/fail, majority pass with some failures (discrepancy), partial pass with mixed results, or other patterns. The engine also tracks metadata including property names, execution times, violation counts, and tool-specific metrics. Results are cached during the session to enable rapid re-comparison with different filters. The implementation includes validation for input format integrity and error handling for malformed JSON or unsupported tool output formats.

WHAT IT CANNOT SEE

This tool cannot analyze contracts that have not been verified or do not have formal verification outputs available from at least two tools. It is limited by the completeness and accuracy of the input verification tool outputs—it cannot fill gaps or validate tool results. It cannot generate new verification results—only compares existing outputs from the tools you provide. It cannot resolve discrepancies between tools without additional manual analysis or re-running verification with different configurations. The tool relies on the accuracy and completeness of source code verification and verification tool data; incomplete or incorrect tool outputs will lead to inaccurate comparisons. It cannot compare tools that use fundamentally different verification approaches or output formats that are not parsed by the engine. The comparison is limited to the specific invariants and properties that were actually checked by the input tools—properties not covered by any tool will not appear in the comparison. It cannot detect verification gaps that were not covered by any of the input tools, nor can it verify that the properties checked are sufficient for correctness.

PLEASE NOTE

This is a read‑only analytics tool—it never submits transactions or signs anything. It is designed for EVM‑compatible chains and works with outputs from major formal verification tools. The comparison results are for informational and audit purposes only and should not be considered a guarantee of contract correctness. Always manually review discrepancies and consult the specific tool documentation for interpretation of results. For comprehensive verification, consider running multiple tools with complementary approaches and reviewing coverage gaps.