author-strategy
Pass
Audited by Gen Agent Trust Hub on Sep 17, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill fetches publication data from the National Center for Biotechnology Information (NCBI) PubMed service via the
Biopythonlibrary. This is a well-known and trusted academic data source. - [INDIRECT_PROMPT_INJECTION]: The skill processes external data (article titles and abstracts) retrieved from PubMed. This content is analyzed to produce strategy and career archetype reports. The skill mitigates risks associated with untrusted data through structured analysis and a mandatory human-in-the-loop disambiguation process.
- Ingestion points:
fetch_pubmed.py(viapubmed_parse.py) retrieves publication metadata and abstracts from the PubMed API. - Boundary markers: The results are structured into Markdown reports (
analysis_report.md,archetype_report.md). - Capability inventory: The skill uses the
Bashtool to execute internal Python scripts,Writeto save findings, andReadto present the final analysis to the user. - Sanitization:
pubmed_parse.pytruncates abstracts to 500 characters, reducing the available surface for potential injection payloads.
Audit Metadata