Skills gtars Audit History
📦

Audit History

gtars - 9 audits

Version comparison

Capability and finding changes across audited versions, newest first.

VersionDateResultReview itemsChange vs previous
v9 LatestJul 23, 2026, 02:34 PM No confirmed findings0No capability change
v8 Jul 8, 2026, 07:03 AM No confirmed findings0No capability change
v7 Jul 5, 2026, 09:37 AM No confirmed findings0 Network access
v6 Jun 29, 2026, 12:50 PM 3 confirmed0Network access Contains scripts
v5 Jan 17, 2026, 12:30 AM No confirmed findings0No capability change
v4 Jan 17, 2026, 12:30 AM No confirmed findings0No capability change
v3 Jan 7, 2026, 01:30 AM No confirmed findings0No capability change
v2 Jan 7, 2026, 01:30 AM No confirmed findings0No capability change
v1 Jan 7, 2026, 01:30 AM No confirmed findings0Baseline

Jun 29, 2026, 12:50 PM

Static analysis reported many command, filesystem, and weak-crypto patterns, but most are Markdown examples for legitimate gtars usage. The remaining risk is medium because the skill encourages CLI execution, local file reads and writes, memory-mapped access, and BEDbase fetch/update workflows that can touch user data or the network.

7
Files scanned
1,297
Lines analyzed
6
Review items
2
False positives ignored

Confirmed security concerns (3)

Medium
CLI workflows can execute local data-processing commands
The skill provides many gtars command examples that read input files and write indexes, coverage tracks, split outputs, and score files. This is legitimate for a genomics tool, but assistants should ask before running commands that modify files or process sensitive datasets.
The command examples are explicit and numerous, but they are documented workflows rather than hidden execution. Risk depends on whether an assistant runs them without user approval or with untrusted paths.
Medium
Network cache workflow can fetch remote BEDbase data
The bbcache documentation includes fetch, list, clear, and update operations for BEDbase.org data. This is expected functionality, but it can contact an external service and alter local cache state.
The file names BEDbase.org and shows fetch and update commands. No exfiltration is present, but network access and cache mutation are clear.
Medium
Large local files may be memory-mapped for performance
The skill recommends streaming and memory-mapped access for large BED and genome files. This is a normal performance feature, but it increases local filesystem and resource-use risk for large or sensitive datasets.
Memory-mapped mode is clearly documented, but it is not malicious by itself. The concern is operational impact and sensitive local data access.
Static false positives ignored (2)

These static matches were dismissed by semantic review or matched schema-only tokens, so they are shown for transparency but do not drive the quality score.

Low
Static command-execution detections are mostly Markdown examples
The analyzer flagged many code fences and CLI snippets as Ruby or shell backtick execution. These are documentation examples, not embedded scripts or automatic execution hooks.
The reviewed locations are fenced Markdown examples with gtars commands. No code in the skill automatically executes them.
Low
Weak-cryptography detections are false positives in refget context
The refget documentation describes sequence digests and states that refget-compliant digests use SHA-512 truncated to 48 bytes. I did not find evidence of MD5, SHA-1, or password hashing guidance.
The digest section explicitly references SHA-512 truncation, which is protocol-specific sequence identification. The surrounding lines discuss validation, not security-sensitive authentication.

Detected Patterns

External CLI invocation examplesFilesystem-intensive genomic workflows
Audited by: codex

Jan 17, 2026, 12:30 AM

This is a legitimate genomic interval analysis toolkit. All 187 static findings are false positives: bash command examples in documentation (misidentified as shell execution), standard cryptographic digests for the GA4GH refget protocol in bioinformatics, memory-mapped file access for efficient large file handling, and system info commands during installation. No malicious code patterns, network exfiltration, credential access, or obfuscation detected.

8
Files scanned
1,560
Lines analyzed
3
Review items
0
False positives ignored
Audited by: claude

Jan 17, 2026, 12:30 AM

This is a legitimate genomic interval analysis toolkit. All 187 static findings are false positives: bash command examples in documentation (misidentified as shell execution), standard cryptographic digests for the GA4GH refget protocol in bioinformatics, memory-mapped file access for efficient large file handling, and system info commands during installation. No malicious code patterns, network exfiltration, credential access, or obfuscation detected.

8
Files scanned
1,560
Lines analyzed
3
Review items
0
False positives ignored
Audited by: claude

Jan 7, 2026, 01:30 AM

This is a legitimate genomic interval analysis toolkit. All documentation describes standard bioinformatics operations. No malicious code patterns, network exfiltration, credential access, or obfuscation detected. The tool performs file I/O for genomic formats only.

7
Files scanned
1,297
Lines analyzed
3
Review items
0
False positives ignored
Audited by: claude

Jan 7, 2026, 01:30 AM

This is a legitimate genomic interval analysis toolkit. All documentation describes standard bioinformatics operations. No malicious code patterns, network exfiltration, credential access, or obfuscation detected. The tool performs file I/O for genomic formats only.

7
Files scanned
1,297
Lines analyzed
3
Review items
0
False positives ignored
Audited by: claude

Jan 7, 2026, 01:30 AM

This is a legitimate genomic interval analysis toolkit. All documentation describes standard bioinformatics operations. No malicious code patterns, network exfiltration, credential access, or obfuscation detected. The tool performs file I/O for genomic formats only.

7
Files scanned
1,297
Lines analyzed
3
Review items
0
False positives ignored
Audited by: claude