Ulysses: A Security Harness For Synthesizing Verifiable Web Exploits

Evidence record

Ulysses: A Security Harness For Synthesizing Verifiable Web Exploits

Read the original source ↗

Claim-level citations (2)

• supportsUlysses verifies agent-generated web exploit claims against execution effects: Ulysses reports four additional previously undocumented proofs of exploitability beyond reproducing ten of the 25 advisories in its chosen WebExploitGym collection; the paper does not equate those proofs with newly assigned CVEs.

  Abstract, WebExploitGym results and four additional proof paths

• supportsUlysses verifies agent-generated web exploit claims against execution effects: Ulysses synthesizes testable web exploit hypotheses and independently checks observable application effects; the authors report 10 successful proof-of-exploitability reproductions among 25 WebExploitGym advisories across five programming languages and seven frameworks.

  Abstract: independently observed execution effects, 25 advisories, 10 verified proofs and four additional proof paths