{"id":"EUVD-2026-114361","enisaUuid":"a857c9c3-5517-3cc6-ae8a-60d1701bf9ea","description":"praktor is a multi-agent Claude Code orchestrator with Telegram input and output, Docker isolation, and a Mission Control UI. Prior to 0.9.1, both Telegram upload sites in `internal/telegram/bot.go` concatenate the attacker-controlled document filename into a destination path without sanitization, and `path.Join` collapses parent-directory traversal. A remote Telegram user who can message the bot can therefore write outside `uploads/` within the persistent agent workspace mounted at `/workspace/agent`. Because `allow_from: []` permits all senders by default, no prior authentication or privilege is required in that configuration. The attacker can plant or overwrite `CLAUDE.md`, instruction, skill, hook, or configuration files that the Claude agent reads and operates on across later runs, turning a document upload into stored instruction injection. The demonstrated arbitrary-write scope is the persistent agent workspace volume, not an unrestricted host-filesystem write. This issue is fixed in version 0.9.1.","datePublished":"Oct 30, 2026, 7:00:25 PM","dateUpdated":"Oct 30, 2026, 7:46:25 PM","baseScore":8.2,"baseScoreVersion":"3.1","baseScoreVector":"CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:L/I:H/A:N","references":"https://github.com/mtzanidakis/praktor/security/advisories/GHSA-48v2-6jcm-fc99\nhttps://github.com/mtzanidakis/praktor/pull/228\nhttps://github.com/mtzanidakis/praktor/commit/c3731eeda01c0b3ec1dbd13cccd5a147c3182141\nhttps://github.com/mtzanidakis/praktor/releases/tag/v0.9.1\n","aliases":"CVE-2026-62331\nGHSA-48v2-6jcm-fc99\n","assigner":"GitHub_M","epss":0.78,"enisaIdProduct":[{"id":"082ba8fe-2f6e-32bc-95f8-fc86860483a5","product":{"name":"praktor","vendor":{"name":"mtzanidakis"}},"product_version":"\u003c 0.9.1"}],"enisaIdVendor":[{"id":"6967a166-7915-3478-b5e6-da07d06a7027","vendor":{"name":"mtzanidakis"}}],"enisaIdVulnerability":[{"id":"5286dfb5-3909-3425-ad44-d85c1bf92eb0","vulnerability":{"id":"CVE-2026-62331","description":"praktor is a multi-agent Claude Code orchestrator with Telegram input and output, Docker isolation, and a Mission Control UI. Prior to 0.9.1, both Telegram upload sites in `internal/telegram/bot.go` concatenate the attacker-controlled document filename into a destination path without sanitization, and `path.Join` collapses parent-directory traversal. A remote Telegram user who can message the bot can therefore write outside `uploads/` within the persistent agent workspace mounted at `/workspace/agent`. Because `allow_from: []` permits all senders by default, no prior authentication or privilege is required in that configuration. The attacker can plant or overwrite `CLAUDE.md`, instruction, skill, hook, or configuration files that the Claude agent reads and operates on across later runs, turning a document upload into stored instruction injection. The demonstrated arbitrary-write scope is the persistent agent workspace volume, not an unrestricted host-filesystem write. This issue is fixed in version 0.9.1.","datePublished":"Oct 30, 2026, 7:00:25 PM","dateUpdated":"Oct 30, 2026, 7:46:25 PM","status":"PUBLISHED","baseScore":8.2,"baseScoreVersion":"3.1","baseScoreVector":"CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:L/I:H/A:N","references":"https://github.com/mtzanidakis/praktor/security/advisories/GHSA-48v2-6jcm-fc99\nhttps://github.com/mtzanidakis/praktor/pull/228\nhttps://github.com/mtzanidakis/praktor/commit/c3731eeda01c0b3ec1dbd13cccd5a147c3182141\nhttps://github.com/mtzanidakis/praktor/releases/tag/v0.9.1\n","enisa_id":"EUVD-2026-114361\n","assigner":"GitHub_M","epss":0.78,"dataProcessed":"Nov 15, 2026, 7:18:47 AM","vulnerabilityProduct":[{"id":"2786db63-583d-3690-bb34-ff6f19ed736d","product":{"name":"praktor","vendor":{"name":"mtzanidakis"}},"product_version":"\u003c 0.9.1"}],"vulnerabilityVendor":[{"id":"1b17d5c5-d96d-3e5a-b2fa-09e31cff8bb9","vendor":{"name":"mtzanidakis"}}]}},{"id":"924d176f-8cdb-3ae2-aef6-818a7ef8d42a","vulnerability":{"id":"GHSA-48v2-6jcm-fc99","description":"praktor: Path traversal in Telegram file uploads allows writing outside the agent workspace uploads/ directory (stored instruction injection)","datePublished":"Nov 3, 2026, 5:27:25 PM","dateUpdated":"Nov 3, 2026, 5:27:25 PM","baseScore":0.0,"references":"https://github.com/mtzanidakis/praktor/security/advisories/GHSA-48v2-6jcm-fc99\nhttps://nvd.nist.gov/vuln/detail/CVE-2026-62331\nhttps://github.com/mtzanidakis/praktor/pull/228\nhttps://github.com/mtzanidakis/praktor/commit/c3731eeda01c0b3ec1dbd13cccd5a147c3182141\nhttps://github.com/mtzanidakis/praktor\nhttps://github.com/mtzanidakis/praktor/releases/tag/v0.9.1\n","enisa_id":"EUVD-2026-114361\n","aliases":"CVE-2026-62331\nGHSA-48v2-6jcm-fc99\n","epss":0.0,"dataProcessed":"Nov 15, 2026, 7:05:40 AM","vulnerabilityProduct":[],"vulnerabilityVendor":[]}}],"enisaIdAdvisory":[]}