Habilidades machine-learning Historial de auditorías
📦

Historial de auditorías

machine-learning - 6 auditorías

Comparación de versiones

Cambios de capacidades y hallazgos entre versiones auditadas, primero las más recientes.

VersiónFechaResultadoElementos de revisiónCambio vs anterior
v6 Más reciente5 jul 2026, 12:49 Sin hallazgos confirmados0Sin cambios de capacidad
v5 5 jul 2026, 12:49 Sin hallazgos confirmados0Comandos externos
v4 27 jun 2026, 18:08 Sin hallazgos confirmados0 Comandos externos
v3 16 ene 2026, 13:17 Sin hallazgos confirmados0Sin cambios de capacidad
v2 16 ene 2026, 13:17 Sin hallazgos confirmados0Comandos externos
v1 10 ene 2026, 09:29 Sin hallazgos confirmados0Base

5 jul 2026, 12:49

All six static findings are false positives caused by Markdown formatting or machine learning terminology. No prompt injection, malicious intent, data exfiltration, or unsafe execution behavior was found in the reviewed files.

2
Archivos escaneados
374
Líneas analizadas
1
Elementos de revisión
0
Falsos positivos ignorados
Auditado por: codex

5 jul 2026, 12:49

All six static findings are false positives caused by Markdown formatting or machine learning terminology. No prompt injection, malicious intent, data exfiltration, or unsafe execution behavior was found in the reviewed files.

2
Archivos escaneados
374
Líneas analizadas
1
Elementos de revisión
0
Falsos positivos ignorados
Auditado por: codex

27 jun 2026, 18:08

Static analysis reported external command, weak cryptography, and reconnaissance patterns. Manual review found instructional Markdown, Python preprocessing examples, and ML terminology, with no executable shell logic or malicious intent.

2
Archivos escaneados
374
Líneas analizadas
0
Elementos de revisión
4
Falsos positivos ignorados
Falsos positivos estáticos ignorados (4)

Estas coincidencias estáticas fueron descartadas por la revisión semántica o coincidieron con tokens solo de esquema, por lo que se muestran por transparencia, pero no afectan la puntuación de calidad.

Bajo
False Positive: Python Examples Flagged as Shell Execution
The flagged references/preprocessing.md locations are fenced Python examples for imputation, scaling, encoding, feature engineering, outlier handling, and pipelines. No shell command execution or command interpolation is present.
The reviewed locations are Markdown code fences containing ordinary Python data science snippets. I found no subprocess, shell, network, or secret handling behavior in those examples.
Bajo
False Positive: Backtick Formatting Flagged as Commands
The SKILL.md locations use Markdown backticks for a model directory tree and reference file names. This is documentation formatting, not command execution.
The flagged text is inside a Markdown code block or inline Markdown formatting. It contains paths and file names, not runnable shell instructions.
Bajo
False Positive: Weak Cryptography Pattern
The SKILL.md description is plain metadata about machine learning workflows. No cryptographic algorithm, hash function, cipher, or credential handling appears at the flagged location.
The flagged location contains only the skill description text. The weak cryptography alert has no semantic support in the reviewed content.
Bajo
False Positive: System Reconnaissance Pattern
The flagged SKILL.md locations contain ML feature and tuning terms, including geospatial features and grid search. They do not request host, network, user, process, or environment reconnaissance.
The surrounding context is machine learning feature engineering and hyperparameter tuning. I found no commands or instructions for gathering system information.
No se registraron hallazgos de seguridad confirmados en esta auditoría completada.
Auditado por: codex

16 ene 2026, 13:17

Pure documentation skill containing only markdown guidance for ML development. All static findings are FALSE POSITIVES: The 'external_commands' detections are Python variable names (X_train, X_test) that regex incorrectly matches as backticks; the 'C2 keywords' and 'weak cryptographic algorithm' detections are SHA256 content hash strings in metadata; the 'system reconnaissance' detections are standard ML monitoring references. No executable code, scripts, network calls, or command execution capabilities exist.

3
Archivos escaneados
564
Líneas analizadas
1
Elementos de revisión
0
Falsos positivos ignorados
Auditado por: claude

16 ene 2026, 13:17

Pure documentation skill containing only markdown guidance for ML development. All static findings are FALSE POSITIVES: The 'external_commands' detections are Python variable names (X_train, X_test) that regex incorrectly matches as backticks; the 'C2 keywords' and 'weak cryptographic algorithm' detections are SHA256 content hash strings in metadata; the 'system reconnaissance' detections are standard ML monitoring references. No executable code, scripts, network calls, or command execution capabilities exist.

3
Archivos escaneados
564
Líneas analizadas
1
Elementos de revisión
0
Falsos positivos ignorados
Auditado por: claude

10 ene 2026, 09:29

Pure documentation skill with no executable code. Contains only markdown guidance for ML development patterns. No scripts, network calls, file system access beyond its own files, or command execution capabilities.

2
Archivos escaneados
374
Líneas analizadas
0
Elementos de revisión
0
Falsos positivos ignorados
No se registraron hallazgos de seguridad confirmados en esta auditoría completada.
Auditado por: claude