cargo package report

Is tokio-tar safe?

1 known vulnerability, worst severity HIGH.

// reach

0 direct dependencies

none carry a known advisory

    0 packages depend on it

    an advisory here reaches each of them

      Create a free accountfor every dependency path, dependent and what to upgrade
      // ai model usage

      Tracked for PyPI packages. HuggingFace models declare Python dependencies, so cargo packages are not covered.


      cvss
      0.0
      high

      severity out of 10

      epss
      0.00%
      medium

      chance of exploitation in 30 days, 51st percentile of all CVEs

      xyz score
      0.0
      low

      CyberXYZ composite out of 10

      fig. 01 — GHSA-j5gw-2vrg-8fgx, the advisory selected below

      // 1 advisories

      GHSA-j5gw-2vrg-8fgx

      HIGHCVE-2025-62518
      // summary

      Versions of astral-tokio-tar prior to 0.5.6 contain a boundary parsing vulnerability that allows attackers to smuggle additional archive entries by exploiting inconsistent PAX/ustar header handling. When processing archives with PAX-extended headers containing size overrides, the parser incorrectly advances stream position based on ustar header size (often zero) instead of the PAX-specified size, causing it to interpret file content as legitimate tar headers.

      This vulnerability was disclosed to multiple Rust tar parsers, all derived from the original async-tar fork of tar-rs.

      // vulnerability description

      The vulnerability stems from inconsistent handling of PAX extended headers versus ustar headers when determining file data boundaries. Specifically:

      • PAX header correctly specifies the file size (e.g., size=1048576)
      • ustar header incorrectly specifies zero size (size=000000000000)
      • tokio-tar advances the stream position based on the ustar size (0 bytes)
      • Inner content is then interpreted as legitimate outer archive entries
      // attack mechanism

      When a TAR file contains:

      • An outer entry with PAX size=N but ustar size=0
      • File data that begins with valid TAR header structures
      • The parser treats inner content as additional outer entries

      This creates a header/data desynchronization where the parser's position becomes misaligned with actual file boundaries.

      // root cause
      // Vulnerable: Uses ustar size instead of PAX override
      let file_size = header.size(); // Returns 0 from ustar field
      let next_pos = current_pos + 512 + pad_to_512(file_size); // Advances 0 bytes
      
      // Fixed: Apply PAX overrides before position calculation
      let mut file_size = header.size();
      if let Some(pax_size) = pending_pax.get("size") {
          file_size = pax_size.parse().unwrap();
      }
      let next_pos = current_pos + 512 + pad_to_512(file_size); // Correct advance
      // impact

      The impact of this vulnerability depends on where astral-tokio-tar is used, and whether it is used to extract untrusted tar archives. If used to extract untrusted inputs, it may result in unexpected attacker-controlled access to the filesystem, in turn potential resulting in arbitrary code execution or credential exfiltration.

      See GHSA-w476-p2h3-79g9 for how this vulnerability affects uv, astral-tokio-tar's primary downstream user. Observe that unlike this advisory, uv's advisory is considered low severity due to overlap with intentional existing capabilities in source distributions.

      // workarounds

      Users are advised to upgrade to version 0.5.6 or newer to address this advisory.

      There is no workaround other than upgrading.

      // timeline

      | Date | Event | | --- | --- | | Aug 21, 2025 | Vulnerability discovered by Edera Security Team | | Aug 21, 2025 | Initial analysis and PoC confirmed | | Aug 22, 2025 | Maintainers notified (privately) | | Aug 25, 2025 | Private patch and test suite shared | | Oct 7, 2025 | Text freeze for GHSA | | Oct 21, 2025 | Coordinated public disclosure and patched releases |

      // credits
      • Discovered by: Steven Noonan (Edera) and Alex Zenla (Edera)
      • Coordinated disclosure: Ann Wallace (Edera)
      // cvss v3.1 vector

      CVSS:3.1/AV:N/AC:L/PR:N/UI:R/S:U/C:H/I:H/A:N

      Attack vector
      Network
      Attack complexity
      Low
      Privileges required
      None
      User interaction
      Required
      Scope
      Unchanged
      Confidentiality
      High
      Integrity
      High
      Availability
      None

      Checked 2026-09-28 at 19:51 UTC. The most recent advisory here was published 2025-10-21. Updated continuously from NVD, GHSA, OSV and CNA feeds.

      Think a verdict here is wrong? Tell us — we respond within 2 business days.