{"id":"CVE-2026-9366","published":"2026-05-24T08:15:09Z","modified":"2026-05-27T17:27:32Z","title":"NousResearch hermes-agent prompt_builder.py _scan_context_content injection","vendor":"NousResearch","product":"hermes-agent","tier":"tail","description":{"text":"A vulnerability was found in NousResearch hermes-agent 2026.4.23. The impacted element is the function _scan_context_content of the file agent/prompt_builder.py. The manipulation results in injection. The attack may be performed from remote. The exploit has been made public and could be used. The vendor was contacted early about this disclosure but did not respond in any way.","provenance":"source","from":"CVE.org"},"cvss":{"score":7.3,"severity":"HIGH","version":"3.1","vector":"CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:L/I:L/A:L/E:P/RL:X/RC:R","provenance":"source","from":"CVE.org CNA"},"epss":{"score":0.00305,"percentile":0.23843,"provenance":"source"},"cwes":[{"id":"CWE-74","name":"Improper Neutralization of Special Elements in Output Used by a Downstream Component"},{"id":"CWE-707","name":"Improper Neutralization"}],"references":[{"url":"https://vuldb.com/vuln/365329","tags":["vdb-entry","technical-description"]},{"url":"https://vuldb.com/vuln/365329/cti","tags":["signature","permissions-required"]},{"url":"https://vuldb.com/submit/812227","tags":["third-party-advisory"]},{"url":"https://gist.github.com/YLChen-007/581fd92de5548fbaacb2092e848a75cc","tags":["exploit"]}],"priority":{"score":14.9,"factors":[{"name":"EPSS percentile","points":4.77},{"name":"Public exploit code","points":5.0},{"name":"CVSS severity","points":5.11}]},"vendorSlug":"nousresearch","affectedVersions":[{"vendor":"NousResearch","product":"hermes-agent","versions":["2026.4.23"]}],"ssvc":{"exploitation":"poc","automatable":"yes","technicalImpact":"partial","timestamp":"2026-05-27T16:25:56.063453Z","from":"CISA ADP","provenance":"source"},"affected":[{"ecosystem":"PyPI","package":"hermes-agent","vulnerableRanges":["<0.15.0"],"patched":["0.15.0"],"from":["GHSA-pgp4-xr4j-h5cg","PYSEC-2026-2514"],"provenance":"source"}],"summary":{"text":"NousResearch hermes-agent contains a injection flaw in the _scan_context_content function of agent/prompt_builder.py, where improper neutralization of special elements allows crafted input to be treated as instructions by a downstream component. A remote attacker needs no privileges or user interaction to exploit it, and the attack requires only network reachability to the affected agent.","provenance":"ai","model":"claude-haiku-5-5","generatedAt":"2026-11-20T21:32:29Z"},"euvd":{"id":"EUVD-2026-31580","exploited":false,"provenance":"source"},"exploitRefs":["https://gist.github.com/YLChen-007/581fd92de5548fbaacb2092e848a75cc"]}