{"id":"GHSA-gpv5-7x3g-ghjv","aliases":[],"status":"published","title":"fast-xml-parser regex vulnerability patch could be improved from a safety perspective","description":"### Summary\nThis is a comment on https://github.com/NaturalIntelligence/fast-xml-parser/security/advisories/GHSA-6w63-h3fj-q4vw and the patches fixing it.\n\n### Details\nThe code which validates a name calls the validator:\nhttps://github.com/NaturalIntelligence/fast-xml-parser/blob/ecf6016f9b48aec1a921e673158be0773d07283e/src/xmlparser/DocTypeReader.js#L145-L153\nThis checks for the presence of an invalid character.  Such an approach is always risky, as it is so easy to forget to include an invalid character in the list.  A safer approach is to validate entity names against the XML specification: https://www.w3.org/TR/xml11/#sec-common-syn - an ENTITY name is a Name:\n\n```\n[4]   NameStartChar ::= \":\" | [A-Z] | \"_\" | [a-z] | [#xC0-#xD6] | [#xD8-#xF6] | [#xF8-#x2FF] | [#x370-#x37D] |\n                        [#x37F-#x1FFF] | [#x200C-#x200D] | [#x2070-#x218F] | [#x2C00-#x2FEF] | [#x3001-#xD7FF] |\n                        [#xF900-#xFDCF] | [#xFDF0-#xFFFD] | [#x10000-#xEFFFF]\n[4a]  NameChar ::= NameStartChar | \"-\" | \".\" | [0-9] | #xB7 | [#x0300-#x036F] | [#x203F-#x2040]\n[5]   Name ::= NameStartChar (NameChar)*\n```\n\nso the safest way to validate an entity name is to build a regex to represent this expression and check whether the name given matches the regex.  (Something along the lines of `/^[name start char class][name char class]*$/`.)  There's probably a nice way to simplify the explicit list rather than typing it out verbatim using Unicode character properties, but I don't know enough to do so.","facts":{"in_kev":false,"known_exploited":false,"public_exploit":false,"fix_available":true},"published":"2023-06-15T19:05:13Z","modified":"2023-11-29T00:28:48Z","fixes":[{"ecosystem":"npm","package":"fast-xml-parser","fixed_in":["4.2.5"],"sources":["ghsa","osv"]}],"affected":[{"source":"ghsa","ecosystem":"npm","package":"fast-xml-parser","ranges":[{"type":"ECOSYSTEM","events":[{"introduced":"4.2.4"},{"last_affected":"4.2.4"},{"fixed":"4.2.5"}]}],"versions":["4.2.4"]},{"source":"osv","ecosystem":"npm","package":"fast-xml-parser","ranges":[{"type":"SEMVER","events":[{"introduced":"4.2.4"},{"fixed":"4.2.5"}]}],"versions":["4.2.4"]}],"references":[{"url":"https://github.com/NaturalIntelligence/fast-xml-parser/security/advisories/GHSA-6w63-h3fj-q4vw","tags":["web"]},{"url":"https://github.com/NaturalIntelligence/fast-xml-parser/security/advisories/GHSA-gpv5-7x3g-ghjv","tags":["web"]},{"url":"https://github.com/NaturalIntelligence/fast-xml-parser/commit/9a880b887916855c3a510869fd1ee268d7fe58b1","tags":["web"]},{"url":"https://github.com/advisories/GHSA-gpv5-7x3g-ghjv"},{"url":"https://github.com/NaturalIntelligence/fast-xml-parser","tags":["package"]}],"sources":{"ghsa":[{"id":"GHSA-gpv5-7x3g-ghjv","title":"fast-xml-parser regex vulnerability patch could be improved from a safety perspective","description":"### Summary\nThis is a comment on https://github.com/NaturalIntelligence/fast-xml-parser/security/advisories/GHSA-6w63-h3fj-q4vw and the patches fixing it.\n\n### Details\nThe code which validates a name calls the validator:\nhttps://github.com/NaturalIntelligence/fast-xml-parser/blob/ecf6016f9b48aec1a921e673158be0773d07283e/src/xmlparser/DocTypeReader.js#L145-L153\nThis checks for the presence of an invalid character.  Such an approach is always risky, as it is so easy to forget to include an invalid character in the list.  A safer approach is to validate entity names against the XML specification: https://www.w3.org/TR/xml11/#sec-common-syn - an ENTITY name is a Name:\n\n```\n[4]   NameStartChar ::= \":\" | [A-Z] | \"_\" | [a-z] | [#xC0-#xD6] | [#xD8-#xF6] | [#xF8-#x2FF] | [#x370-#x37D] |\n                        [#x37F-#x1FFF] | [#x200C-#x200D] | [#x2070-#x218F] | [#x2C00-#x2FEF] | [#x3001-#xD7FF] |\n                        [#xF900-#xFDCF] | [#xFDF0-#xFFFD] | [#x10000-#xEFFFF]\n[4a]  NameChar ::= NameStartChar | \"-\" | \".\" | [0-9] | #xB7 | [#x0300-#x036F] | [#x203F-#x2040]\n[5]   Name ::= NameStartChar (NameChar)*\n```\n\nso the safest way to validate an entity name is to build a regex to represent this expression and check whether the name given matches the regex.  (Something along the lines of `/^[name start char class][name char class]*$/`.)  There's probably a nice way to simplify the explicit list rather than typing it out verbatim using Unicode character properties, but I don't know enough to do so.","status":"published","published":"2023-06-15T19:05:13Z","modified":"2023-11-29T00:28:48Z","updated_at":"2026-11-02T14:37:22.840002Z"}],"osv":[{"id":"GHSA-gpv5-7x3g-ghjv","title":"fast-xml-parser regex vulnerability patch could be improved from a safety perspective","description":"### Summary\nThis is a comment on https://github.com/NaturalIntelligence/fast-xml-parser/security/advisories/GHSA-6w63-h3fj-q4vw and the patches fixing it.\n\n### Details\nThe code which validates a name calls the validator:\nhttps://github.com/NaturalIntelligence/fast-xml-parser/blob/ecf6016f9b48aec1a921e673158be0773d07283e/src/xmlparser/DocTypeReader.js#L145-L153\nThis checks for the presence of an invalid character.  Such an approach is always risky, as it is so easy to forget to include an invalid character in the list.  A safer approach is to validate entity names against the XML specification: https://www.w3.org/TR/xml11/#sec-common-syn - an ENTITY name is a Name:\n\n```\n[4]   NameStartChar ::= \":\" | [A-Z] | \"_\" | [a-z] | [#xC0-#xD6] | [#xD8-#xF6] | [#xF8-#x2FF] | [#x370-#x37D] |\n                        [#x37F-#x1FFF] | [#x200C-#x200D] | [#x2070-#x218F] | [#x2C00-#x2FEF] | [#x3001-#xD7FF] |\n                        [#xF900-#xFDCF] | [#xFDF0-#xFFFD] | [#x10000-#xEFFFF]\n[4a]  NameChar ::= NameStartChar | \"-\" | \".\" | [0-9] | #xB7 | [#x0300-#x036F] | [#x203F-#x2040]\n[5]   Name ::= NameStartChar (NameChar)*\n```\n\nso the safest way to validate an entity name is to build a regex to represent this expression and check whether the name given matches the regex.  (Something along the lines of `/^[name start char class][name char class]*$/`.)  There's probably a nice way to simplify the explicit list rather than typing it out verbatim using Unicode character properties, but I don't know enough to do so.","status":"published","published":"2023-06-15T19:05:13Z","modified":"2023-06-15T19:05:13Z","updated_at":"2026-11-02T15:46:48.191583Z"}]}}