Most of the bugs Claude Mythos found have never been checked by a human
Echo's analysis found only 1,900 of 23,019 Claude Mythos-found vulnerabilities were externally reviewed, 90.8% held up, but the model overstated most severities.
Echo analyzed results from Anthropic's Claude Mythos Preview vulnerability sweep across 281 open-source projects, which produced 23,019 candidate vulnerabilities, of which only 1,900 were externally reviewed. Of those, 90.8% held up as real, 1,451 of 1,596 maintainer reports were acknowledged, 97 fixes landed upstream, and 88 became advisories, but 14 of the 27 CVE-assigned severity ratings mismatched independent scoring, mostly overstated. On Anthropic's SpiderMonkey benchmark, Claude Mythos turned known crashes into working code execution exploits in 72.4% of 250 trials, versus below 1% for Claude Opus 4.6. Echo cautions the reviewed sample likely was not randomly drawn, so the accuracy figure may not generalize to the other 21,119 unreviewed candidates.
- 23,019 candidate vulnerabilities across 281 open-source projects; only 1,900 externally reviewed, 90.8% confirmed as real.
- 1,451 of 1,596 maintainer reports acknowledged; 97 fixes landed upstream and 88 became published advisories.
- Of 27 CVE-assigned findings, 14 carried severity mismatches: 13 overstated by Mythos, one understated (jq).
- Claude Mythos turned SpiderMonkey crashes into code execution in 72.4% of 250 trials; Claude Opus 4.6 below 1%.
- A Linux kernel use-after-free was chained into a root exploit for under $2,000 in inference costs.
- Vendors
- Anthropic
- AI models
- Claude MythosClaude Opus 4.6
From the original article
3 min read at helpnetsecurity.com
Anthropic pointed Claude Mythos Preview at 281 open-source projects and collected 23,019 candidate vulnerabilities. External security firms reviewed 1,900 of them. Maintainers received 1,596 reports and acknowledged 1,451; 97 fixes landed upstream, and 88 findings became published security advisories, with counts current as of May 22, 2026.
By @helpnetsecurity · helpnetsecurity.com
A short excerpt; the article and its images belong to the publisher. Summary, key facts and assessment above are AI-written from the article and can be wrong.