Practical Picks PP-285: earlier hands-on check record, October 3, 2026 Observed 2026-10-03, Linux. All inputs were synthetic or public RFC test data. Historical record. This small check ran on October 3, 2026, before Practical Picks adopted its research-only method later that day. It is kept as dated, bounded evidence. It is not current testing, and it did not decide the article's recommendations, which since October 20, 2026 rest on public sources. Practical Picks no longer runs product checks; the acquisition and replay instructions below describe how this historical record was produced. Notes, October 20, 2026: - OATH Toolkit 2.6.15 (October 13, 2026) is now the current upstream release. The Ubuntu 24.04 package checked here, 2.6.11-2.1ubuntu0.1, is the revision Ubuntu lists as fixed for CVE-2024-47191 (USN-7059-1). Upstream's current source code also ends a run with exit status 1 when Base32 decoding of the key fails, matching the result recorded below. - The yitsushi/totp-cli 1.9.2 source code prints "Error: ..." to standard output and exits 0 for an invalid seed in `instant`, matching the result recorded below. A source change merged on May 15, 2026 (pull request 150) returns the error instead; no release contained it on October 20, 2026. - The article no longer names or evaluates the 2024 newcomer screened below. No account, real secret, login, hosted call or system installation was used. Question: If a reader already holds a Base32 TOTP seed, which local CLI can generate a code without enrolling another account or vault? A generated TOTP code is not a provider-issued recovery code. These tools cannot recover a lost seed, reset an account or bypass a provider's authentication rules. Fixture: RFC 6238 SHA-1 seed GEZDGNBVGY3TQOJQGEZDGNBVGY3TQOJQ; period 30s. Pinned RFC times and expected 8-digit codes: epoch 59 94287082 epoch 1111111109 07081804 Malformed seed: %%%invalid%%%. Versions: oathtool 2.6.11-2.1ubuntu0.1 and liboath from Ubuntu 24.04 updated packages; yitsushi/totp-cli 1.9.2 Linux x86_64; dotp v0.0.1.20240718.01 Linux amd64. Existing Python 3 was used for the independent HMAC reference. All product files were extracted to /tmp/pp285, not installed system-wide. Combined downloads and extraction occupied 8.6 MB. Exact reproducible acquisition on Linux x86_64 with Python 3 and dpkg-deb: Download /assets/pp285/acquire.py and /assets/pp285/replay.py to the same disposable working directory. python3 acquire.py /tmp/pp285-my-replay python3 replay.py /tmp/pp285-my-replay/unpack/usr/bin/oathtool \ /tmp/pp285-my-replay/unpack/usr/lib/x86_64-linux-gnu \ /tmp/pp285-my-replay/totp-cli /tmp/pp285-my-replay/dotp acquire.py refuses an existing output directory. It downloads from these exact publisher distribution URLs, verifies the SHA-256 values below before use, extracts the two .deb packages with dpkg-deb -x, extracts only the CLI from the totp-cli archive, and marks the two standalone binaries executable. It does not install system packages or submit a seed. The tested official URLs are: https://archive.ubuntu.com/ubuntu/pool/universe/o/oath-toolkit/oathtool_2.6.11-2.1ubuntu0.1_amd64.deb https://archive.ubuntu.com/ubuntu/pool/main/o/oath-toolkit/liboath0t64_2.6.11-2.1ubuntu0.1_amd64.deb https://github.com/yitsushi/totp-cli/releases/download/v1.9.2/totp-cli_Linux_x86_64.tar.gz https://github.com/petlack/dotp/releases/download/v0.0.1.20240718.01/dotp-linux-amd64 Tested download SHA-256 (local provenance hashes, not publisher signatures): oathtool .deb 79f4a92ec4b12a54b0ddf3f3af2cc8f9d12dde4f48bbaaeb42f919f9aab9291b liboath .deb 876b35bf0c62f1f9c4ccb3f88f75c42c489a04823e95504e37e0ce410d169166 yitsushi archive b29673877451bc96d89d99fccaeccc4fe0edccb14ced05e28d293521ce17f1a4 dotp binary b0d73f3222fbb17c1beb20625b43fc1ac7c66de5d97cdd6644d6aba3fafeeee2 Exact tested product command forms (the public replay uses absolute paths): RFC_SEED='GEZDGNBVGY3TQOJQGEZDGNBVGY3TQOJQ' printf '%s\n' "$RFC_SEED" | oathtool --totp=SHA1 --base32 --now=@59 --digits=8 - printf '%s\n' "$RFC_SEED" | oathtool --totp=SHA1 --base32 --now=@1111111109 --digits=8 - printf '%s\n' "$RFC_SEED" | oathtool --totp=SHA1 --base32 - printf '%s\n' "$RFC_SEED" | totp-cli instant printf '%s\n' "$RFC_SEED" | dotp get --secret-stdin The replay repeats each stdin command with `%%%invalid%%%`, records exit, stdout and stderr, then retries the valid seed. It also checks inherited file descriptors for dotp and oathtool (via Linux /proc/self/fd/N). The command list above illustrates stdin use; do not paste a real seed into shell history. The replay passes the public test seed directly to subprocess stdin, with an isolated temporary HOME. It also passes the seed through an inherited file descriptor to dotp without putting it in argv. Oathtool also reads an inherited pipe through its @FILE syntax with /proc/self/fd/N on the tested Linux system. Plain stdin avoids argv exposure for both tools. Run the public replay after acquiring the three tools at disposable local paths, using the tested acquisition command above or equivalent verified files: python3 replay.py ABSOLUTE_OATHTOOL ABSOLUTE_LIBOATH_DIR ABSOLUTE_TOTP_CLI ABSOLUTE_DOTP Download the script from /assets/pp285/replay.py. It requires no Python packages. It creates a temporary HOME, checks that no files appeared there, and writes replay-results.json beside the script. The retained run output is /assets/pp285/replay-results.json. That JSON includes each full product argv, exit code, stdout, stderr and before/after timestamp. Replace the task-local paths with your own when rerunning. Observed result: oathtool produced both pinned RFC codes exactly. All three tools produced the same 6-digit value as independent HMAC in a captured live 30-second period. The replay rejects a run crossing the period boundary. yitsushi and dotp have no timestamp override in the tested direct-code commands, so their live-window checks are not pinned historical-time tests. Both fd routes also matched. On malformed input, oathtool and dotp exited 1 with empty stdout; yitsushi exited 0 and wrote an error message to stdout. A subsequent valid call passed for each product. This test does not measure vault encryption, safe long-term seed storage, cross-platform behavior, clock skew, live service acceptance, support, source signatures or security of a real seed. TOTP codes are not phishing resistant. Do not use an invented test seed for an actual account. Primary sources: https://www.rfc-editor.org/rfc/rfc6238 https://www.nongnu.org/oath-toolkit/ https://www.nongnu.org/oath-toolkit/man-oathtool.html https://ubuntu.com/security/notices/USN-7059-1 https://oath-toolkit.codeberg.page/CVE-2024-47191.html https://github.com/yitsushi/totp-cli/releases/tag/v1.9.2 https://github.com/petlack/dotp/releases/tag/v0.0.1.20240718.01 https://docs.github.com/en/authentication/securing-your-account-with-two-factor-authentication-2fa/configuring-two-factor-authentication-recovery-methods