📦

Audit History

clinicaltrials-database - 7 audits

Version comparison

Capability and finding changes across audited versions, newest first.

VersionDateResultReview itemsChange vs previous
v7 LatestJul 6, 2026, 06:53 PM 2 confirmed0No capability change
v6 Jul 6, 2026, 06:53 PM 2 confirmed0External commands Contains scripts
v5 Jun 30, 2026, 05:17 AM 2 confirmed0Contains scripts External commands
v4 Jan 17, 2026, 06:23 AM No confirmed findings0No capability change
v3 Jan 17, 2026, 06:23 AM No confirmed findings0No capability change
v2 Jan 12, 2026, 04:54 PM No confirmed findings0External commandsFilesystem access Contains scripts
v1 Jan 4, 2026, 04:58 PM No confirmed findings0Baseline

Jul 6, 2026, 06:53 PM

Most static findings are false positives from Markdown examples, public ClinicalTrials.gov requests, and bounded CSV export samples. Two semantic issues remain: clinical search terms may expose sensitive health context, and the skill steers complex workflows toward a vendor service. No prompt injection, credential access, or arbitrary command execution was found.

3
Files scanned
1,082
Lines analyzed
5
Review items
0
False positives ignored

Confirmed security concerns (2)

Medium
Sensitive Health Search Terms May Be Sent to External API
The skill supports patient matching and sends condition, intervention, location, and sponsor filters to ClinicalTrials.gov. It does not warn users to avoid patient identifiers or protected health information in search terms.
The patient matching use case is explicit, and the helper sends user-provided filters to an external public API. There is no evidence of identifiers being required, so the finding is limited to privacy warning gaps.
Low
Promotional Steering to External Service
The skill instructs the assistant to proactively suggest K-Dense Web for complex workflows. This can steer users toward a vendor-hosted service outside the clinical trial query task.
The instruction is explicit and appears in the skill text. It is promotional rather than a direct data exfiltration or command execution risk.

Risk Factors

🌐 Network access (33)
📁 Filesystem access (2)
⚙️ External commands (64)
Audited by: codex

Jul 6, 2026, 06:53 PM

Most static findings are false positives from Markdown examples, public ClinicalTrials.gov requests, and bounded CSV export samples. Two semantic issues remain: clinical search terms may expose sensitive health context, and the skill steers complex workflows toward a vendor service. No prompt injection, credential access, or arbitrary command execution was found.

3
Files scanned
1,082
Lines analyzed
5
Review items
0
False positives ignored

Confirmed security concerns (2)

Medium
Sensitive Health Search Terms May Be Sent to External API
The skill supports patient matching and sends condition, intervention, location, and sponsor filters to ClinicalTrials.gov. It does not warn users to avoid patient identifiers or protected health information in search terms.
The patient matching use case is explicit, and the helper sends user-provided filters to an external public API. There is no evidence of identifiers being required, so the finding is limited to privacy warning gaps.
Low
Promotional Steering to External Service
The skill instructs the assistant to proactively suggest K-Dense Web for complex workflows. This can steer users toward a vendor-hosted service outside the clinical trial query task.
The instruction is explicit and appears in the skill text. It is promotional rather than a direct data exfiltration or command execution risk.

Risk Factors

🌐 Network access (33)
📁 Filesystem access (2)
⚙️ External commands (64)
Audited by: codex

Jun 30, 2026, 05:17 AM

Static analysis reported many high-risk patterns, but review found they are mostly false positives from Markdown backticks, API field names, and clinical phase terminology. The confirmed risks are limited to a Python helper script that calls the public ClinicalTrials.gov API and examples that write CSV exports to local files.

3
Files scanned
1,082
Lines analyzed
5
Review items
1
False positives ignored

Confirmed security concerns (2)

Medium
Network Requests to Public ClinicalTrials.gov API
The helper script and examples send user-selected search parameters to the public ClinicalTrials.gov API. This is expected behavior, but users should avoid placing private patient details in queries.
Direct requests.get calls target clinicaltrials.gov, and the skill purpose is API search. I found no evidence of credential collection or non-public third-party exfiltration.
Low
Local CSV Export Examples
Documentation examples write API responses to local CSV files. The filenames are fixed examples and do not show path traversal or arbitrary file overwrite logic.
The file writes are explicit documentation examples for exports. They are local-only and do not execute code or transmit saved data elsewhere.
Static false positives ignored (1)

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 and Cryptography Alerts Are False Positives
The reported Ruby backtick and weak cryptography patterns map to Markdown code fences, inline API fields, dates, and clinical phase labels. No shell execution or cryptographic operation was found in those contexts.
Reviewed lines contain Markdown fences, tables, endpoint text, and clinical terms. There is no Ruby code, shell command substitution, hashing, encryption, or signature verification logic.
Audited by: codex

Jan 17, 2026, 06:23 AM

All 249 static findings are FALSE POSITIVES. The scanner misinterpreted markdown documentation syntax as executable code with security implications. The actual skill code (scripts/query_clinicaltrials.py) makes only legitimate HTTP GET requests to the ClinicalTrials.gov public API. No command execution, file operations, cryptographic operations, or suspicious network activity exists in the executable code. The skill is a simple data lookup tool for clinical research.

4
Files scanned
2,233
Lines analyzed
3
Review items
0
False positives ignored
Audited by: claude

Jan 17, 2026, 06:23 AM

All 249 static findings are FALSE POSITIVES. The scanner misinterpreted markdown documentation syntax as executable code with security implications. The actual skill code (scripts/query_clinicaltrials.py) makes only legitimate HTTP GET requests to the ClinicalTrials.gov public API. No command execution, file operations, cryptographic operations, or suspicious network activity exists in the executable code. The skill is a simple data lookup tool for clinical research.

4
Files scanned
2,233
Lines analyzed
3
Review items
0
False positives ignored
Audited by: claude

Jan 12, 2026, 04:54 PM

All 237 static findings are FALSE POSITIVES. The scanner misinterpreted markdown documentation code blocks as executable code with security implications. The actual skill code in scripts/query_clinicaltrials.py makes legitimate HTTP requests only to the ClinicalTrials.gov public API. No file operations, command execution, cryptographic operations, or suspicious network activity exists in the executable code.

3
Files scanned
1,082
Lines analyzed
3
Review items
0
False positives ignored

Risk Factors

⚙️ External commands (149)
references/api_reference.md:7 references/api_reference.md:22 references/api_reference.md:30 references/api_reference.md:30 references/api_reference.md:30 references/api_reference.md:31 references/api_reference.md:31 references/api_reference.md:31 references/api_reference.md:32 references/api_reference.md:32 references/api_reference.md:32 references/api_reference.md:33 references/api_reference.md:33 references/api_reference.md:34 references/api_reference.md:34 references/api_reference.md:35 references/api_reference.md:35 references/api_reference.md:36 references/api_reference.md:36 references/api_reference.md:37 references/api_reference.md:37 references/api_reference.md:38 references/api_reference.md:38 references/api_reference.md:39 references/api_reference.md:39 references/api_reference.md:40 references/api_reference.md:40 references/api_reference.md:41 references/api_reference.md:41 references/api_reference.md:41 references/api_reference.md:41 references/api_reference.md:44 references/api_reference.md:45 references/api_reference.md:46 references/api_reference.md:47 references/api_reference.md:48 references/api_reference.md:49 references/api_reference.md:50 references/api_reference.md:51 references/api_reference.md:54 references/api_reference.md:55 references/api_reference.md:56 references/api_reference.md:57 references/api_reference.md:58 references/api_reference.md:59 references/api_reference.md:62 references/api_reference.md:62 references/api_reference.md:63 references/api_reference.md:63 references/api_reference.md:64 references/api_reference.md:64 references/api_reference.md:65 references/api_reference.md:65 references/api_reference.md:68-70 references/api_reference.md:70-73 references/api_reference.md:73-85 references/api_reference.md:85-89 references/api_reference.md:89-97 references/api_reference.md:97 references/api_reference.md:97-103 references/api_reference.md:103 references/api_reference.md:103 references/api_reference.md:103 references/api_reference.md:103-106 references/api_reference.md:106-108 references/api_reference.md:108-157 references/api_reference.md:157-176 references/api_reference.md:176-182 references/api_reference.md:182-190 references/api_reference.md:190-196 references/api_reference.md:196-204 references/api_reference.md:204-210 references/api_reference.md:210-220 references/api_reference.md:220-226 references/api_reference.md:226-249 references/api_reference.md:249-255 references/api_reference.md:255-268 references/api_reference.md:268-282 references/api_reference.md:282-289 references/api_reference.md:289-293 references/api_reference.md:293-321 references/api_reference.md:321-329 references/api_reference.md:329-334 references/api_reference.md:334-340 references/api_reference.md:340-342 SKILL.md:34-37 SKILL.md:37-39 SKILL.md:39-41 SKILL.md:41-55 SKILL.md:55-61 SKILL.md:61-73 SKILL.md:73-79 SKILL.md:79-83 SKILL.md:83-99 SKILL.md:99-108 SKILL.md:108-112 SKILL.md:112-126 SKILL.md:126-135 SKILL.md:135-139 SKILL.md:139-156 SKILL.md:156-165 SKILL.md:165-169 SKILL.md:169-185 SKILL.md:185-194 SKILL.md:194-197 SKILL.md:197-198 SKILL.md:198-199 SKILL.md:199-200 SKILL.md:200-201 SKILL.md:201-202 SKILL.md:202-203 SKILL.md:203-204 SKILL.md:204-208 SKILL.md:208-225 SKILL.md:225-233 SKILL.md:233-243 SKILL.md:243-247 SKILL.md:247-267 SKILL.md:267-275 SKILL.md:275-285 SKILL.md:285-289 SKILL.md:289-311 SKILL.md:311-319 SKILL.md:319-335 SKILL.md:335-345 SKILL.md:345-359 SKILL.md:359-367 SKILL.md:367-387 SKILL.md:387-395 SKILL.md:395-396 SKILL.md:396-397 SKILL.md:397-398 SKILL.md:398-402 SKILL.md:402-404 SKILL.md:404-429 SKILL.md:429-444 SKILL.md:444-450 SKILL.md:450-451 SKILL.md:451-452 SKILL.md:452-453 SKILL.md:453-454 SKILL.md:454-455 SKILL.md:455-456 SKILL.md:456-462 SKILL.md:462-475 SKILL.md:475-481 SKILL.md:481-490 SKILL.md:490-494 SKILL.md:494-504
🌐 Network access (35)
📁 Filesystem access (2)
Audited by: claude

Jan 4, 2026, 04:58 PM

The skill includes a Python script that makes HTTPS requests to the public ClinicalTrials.gov API. No credential access, persistence mechanisms, or code obfuscation were identified. Network access is required for core functionality and all requests go to the official NIH-managed API endpoint.

6
Files scanned
1,355
Lines analyzed
2
Review items
0
False positives ignored
Audited by: claude