bio-proteomics-proteomics-qc
Pass
Audited by Gen Agent Trust Hub on Oct 8, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes external proteomics datasets (intensity matrices and search engine output tables) which serves as a potential vector for indirect prompt injection attacks.
- Ingestion points: Tabular data files (e.g., MaxQuant
proteinGroups.txt, FragPipe TSVs) loaded intopandasDataFrames as described inSKILL.mdandexamples/qc_analysis.py. - Boundary markers: No explicit delimiters or instructions to ignore embedded content are present when processing input data.
- Capability inventory: The skill's capabilities are limited to statistical analysis, data filtering, and dimensionality reduction using
pandas,scipy, andscikit-learn. No high-risk capabilities such as network exfiltration, arbitrary file writing, or shell command execution were identified. - Sanitization: No explicit data validation or escaping mechanisms are implemented for the ingested tabular content.
Audit Metadata