retraction-watcher
Pass
Audited by Gen Agent Trust Hub on Sep 23, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill connects to legitimate academic services including Crossref's API, the National Center for Biotechnology Information (NCBI) PubMed E-utilities, and Open Retractions to verify the status of research citations.
- [COMMAND_EXECUTION]: The skill provides commands to execute a local Python script (
scripts/main.py) to process manuscripts, bibliographies, and reference lists. - [INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data from various sources, representing a potential surface for indirect prompt injection.
- Ingestion points: External content is ingested via file paths (PDF, BibTeX, text) and user-provided URLs in
scripts/main.py. - Boundary markers: The skill does not explicitly define prompt boundaries for the content extracted from processed documents.
- Capability inventory: The skill performs file system reads and HTTPS network requests to academic APIs.
- Sanitization: The script uses regex patterns to extract specific identifiers (DOIs and PMIDs) and performs JSON parsing on API responses, which limits the exposure to malicious input.
Audit Metadata