Skills clinical-reports Audit History
📦

Audit History

clinical-reports - 7 audits

Version comparison

Capability and finding changes across audited versions, newest first.

VersionDateResultReview itemsChange vs previous
v7 LatestJul 5, 2026, 06:46 PM 2 confirmed0No capability change
v6 Jul 5, 2026, 06:46 PM 2 confirmed0 Contains scriptsNetwork access
v5 Jun 30, 2026, 06:24 AM No confirmed findings2Contains scriptsNetwork access
v4 Jan 17, 2026, 06:21 AM No confirmed findings0No capability change
v3 Jan 17, 2026, 06:21 AM No confirmed findings0External commands
v2 Jan 12, 2026, 04:52 PM No confirmed findings0 Contains scripts
v1 Jan 4, 2026, 04:56 PM No confirmed findings0Baseline

Jul 5, 2026, 06:46 PM

Most static findings are false positives caused by clinical terminology, Markdown backticks, and expected local report-output scripts. No evidence found of command injection, credential access, network calls, or hidden data exfiltration in the reviewed code. Two semantic risks remain: an included self-audit that claims safety and a mandatory external AI figure workflow that could expose clinical data.

31
Files scanned
11,881
Lines analyzed
4
Review items
0
False positives ignored

Confirmed security concerns (2)

High
Prompt Injection Attempt Detected
evaluation_output.json claims "no security concerns" and says "all static findings are false positives." This untrusted self-audit can bias automated reviewers and should not ship with the skill.
The file contains a direct safety claim about the package rather than operational code. It may be leftover evaluation output, but it is still untrusted content that can influence review.
High
External AI Figure Workflow May Expose Clinical Data
SKILL.md mandates an AI-generated figure for every clinical report and directs users to a scientific-schematics workflow. Clinical report figure prompts may include patient timelines or trial data, creating a privacy exposure risk if not de-identified.
The text explicitly makes external AI figure generation mandatory and gives a workflow for clinical report diagrams. The privacy risk is inferred from the clinical-report domain and the referenced patient timelines.
Audited by: codex

Jul 5, 2026, 06:46 PM

Most static findings are false positives caused by clinical terminology, Markdown backticks, and expected local report-output scripts. No evidence found of command injection, credential access, network calls, or hidden data exfiltration in the reviewed code. Two semantic risks remain: an included self-audit that claims safety and a mandatory external AI figure workflow that could expose clinical data.

31
Files scanned
11,881
Lines analyzed
4
Review items
0
False positives ignored

Confirmed security concerns (2)

High
Prompt Injection Attempt Detected
evaluation_output.json claims "no security concerns" and says "all static findings are false positives." This untrusted self-audit can bias automated reviewers and should not ship with the skill.
The file contains a direct safety claim about the package rather than operational code. It may be leftover evaluation output, but it is still untrusted content that can influence review.
High
External AI Figure Workflow May Expose Clinical Data
SKILL.md mandates an AI-generated figure for every clinical report and directs users to a scientific-schematics workflow. Clinical report figure prompts may include patient timelines or trial data, creating a privacy exposure risk if not de-identified.
The text explicitly makes external AI figure generation mandatory and gives a workflow for clinical report diagrams. The privacy risk is inferred from the clinical-report domain and the referenced patient timelines.
Audited by: codex

Jun 30, 2026, 06:24 AM

Static analysis reported 569 potential issues, but sampled context shows the weak crypto, reconnaissance, credential, and Ruby backtick alerts are mainly clinical terminology, Markdown formatting, and reference links. No prompt injection, subprocess execution, network client, environment secret access, or malicious exfiltration intent was found. Risk remains medium because the skill is a community healthcare skill with Bash permission and local scripts that read and write clinical documents that may contain PHI.

31
Files scanned
11,881
Lines analyzed
6
Review items
1
False positives ignored
Capability review items (2)

These are real local capabilities that may be expected for this skill, so they require review but are not counted as confirmed malicious behavior.

Medium
Local PHI File Processing
The included scripts read user-supplied clinical reports and can write JSON or Markdown outputs. This is legitimate validation behavior, but users may process protected health information, so outputs must remain in approved local storage.
The file reads and writes are explicit and operate on clinical documents. I did not find network transfer or credential access, so the concern is data handling rather than malicious behavior.
Medium
Broad Tool Permission With Bash Guidance
The skill declares Bash, Write, and Edit permissions and includes a documented bash command for figure generation. The command appears instructional, but broad permissions increase review requirements for a community healthcare skill.
The permission declaration and command block are direct evidence. I did not find subprocess calls in the Python scripts, so legitimate use is likely.
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 Alerts Are Mostly Documentation False Positives
The weak cryptography, reconnaissance, credential, and Ruby backtick detections are explained by clinical abbreviations, HIPAA terminology, Markdown code formatting, and reference URLs. No evidence found of cryptographic implementation misuse, system probing, or credential collection.
Targeted review confirmed the flagged patterns occur in documentation and local utility context. The finding is retained as low because the static analyzer produced many alerts that marketplace reviewers may want summarized.
Audited by: codex

Jan 17, 2026, 06:21 AM

This is a legitimate clinical documentation skill with no security concerns. All static findings are false positives caused by the scanner misinterpreting markdown documentation content. The 'weak cryptographic algorithm' detections are legitimate mentions of HIPAA compliance terms in documentation. The 'Ruby/shell backtick execution' detections are markdown code formatting syntax (`) not actual shell commands. The Python scripts are safe utilities that read files, parse regex patterns, and generate compliance reports with no command injection, credential access, or data exfiltration.

32
Files scanned
12,249
Lines analyzed
2
Review items
0
False positives ignored

Risk Factors

⚙️ External commands (1)
📁 Filesystem access (1)
Audited by: claude

Jan 17, 2026, 06:21 AM

This is a legitimate clinical documentation skill with no security concerns. All static findings are false positives caused by the scanner misinterpreting markdown documentation content. The 'weak cryptographic algorithm' detections are legitimate mentions of HIPAA compliance terms in documentation. The 'Ruby/shell backtick execution' detections are markdown code formatting syntax (`) not actual shell commands. The Python scripts are safe utilities that read files, parse regex patterns, and generate compliance reports with no command injection, credential access, or data exfiltration.

32
Files scanned
12,249
Lines analyzed
2
Review items
0
False positives ignored

Risk Factors

⚙️ External commands (1)
📁 Filesystem access (1)
Audited by: claude

Jan 12, 2026, 04:52 PM

All 560 static analysis findings are false positives. The skill contains no actual security vulnerabilities - it uses no cryptographic algorithms, executes no shell commands, and has no network operations. The tools are legitimate clinical documentation utilities that help ensure HIPAA compliance and proper medical reporting.

30
Files scanned
11,651
Lines analyzed
1
Review items
0
False positives ignored

Risk Factors

📁 Filesystem access (1)
Audited by: claude

Jan 4, 2026, 04:56 PM

This skill includes local Python scripts that read user-provided files and write local outputs. No network access, environment variable access, or external command execution found. Scripts perform legitimate clinical report validation and formatting tasks.

31
Files scanned
11,768
Lines analyzed
2
Review items
0
False positives ignored
Audited by: claude