Online Analysis
PassAudited by ClawScan on May 1, 2026.
Overview
The artifacts show a local, user-directed log/data analysis skill with no evidence of hidden exfiltration or destructive behavior, but with minor packaging and data-handling points to review.
This skill appears safe for local, user-directed testing analysis. Before installing, confirm you trust the publisher, install numpy from a trusted source, run the scripts only on intended log/data files, and treat generated reports as potentially sensitive. Be aware that some documented helper files are not included in the reviewed package.
Findings (4)
Artifact-based informational review of SKILL.md, metadata, install specs, static scan signals, and capability signals. ClawScan does not execute the skill or run runtime probes.
Some advertised features may not work as documented, and users should not substitute or download missing scripts from unreviewed sources without checking them.
The documentation references helper scripts that are not included in the supplied file manifest. This looks like an incomplete package/documentation mismatch rather than malicious behavior, because the included scripts are coherent and no automatic fetching of missing code is instructed.
- `pattern_recognizer.py`: Machine learning model for identifying testing scenarios - `test_case_generator.py`: Automatically generate test cases from extracted rules
Install and run only the files included with the skill unless the publisher provides reviewed replacements for the missing components.
The user is responsible for installing the dependency from a trusted Python package source.
The skill relies on a manual Python package installation that is not captured by an install spec. Numpy is a normal dependency for this purpose, but the install is user-managed and not locked to an exact package artifact.
### Dependencies ```bash pip install numpy ```
Use a trusted package index or controlled environment when installing numpy, and avoid running the skill in sensitive environments if dependencies are not managed.
Running the examples executes local code and may create report files containing derived information from the input data.
The skill documents local Python script execution against user-selected files. This is expected for the stated analysis purpose and is not hidden or automatic.
python scripts/rule_extractor.py transaction_logs.txt > extracted_rules.md python scripts/anomaly_detector.py performance_metrics.json > anomaly_report.md
Run the scripts only on files you intend to analyze, and review generated reports before sharing them.
Generated rule documents can preserve pieces of private log or business data and may be reused by the agent or shared by the user.
The rule extractor can place observed field values from input data into generated markdown rules. This is purpose-aligned, but reports may contain sensitive or misleading values from logs or structured data.
'condition': f"{field} must be one of: {', '.join(map(str, stats['values']))}"Treat outputs as potentially sensitive, sanitize logs before analysis when needed, and do not blindly treat extracted rules as authoritative without review.