homer-motif-analysis
Pass
Audited by Gen Agent Trust Hub on Oct 14, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill installs the HOMER suite via the Bioconda channel and standard data science libraries (pandas, matplotlib, seaborn) via pip. It also uses the utility script
installGenome.plto download large genomic sequence files from the official HOMER servers at UCSD. These are legitimate operations from trusted community sources. - [COMMAND_EXECUTION]: The skill executes various shell commands including Perl scripts (
findMotifsGenome.pl,annotatePeaks.pl), text processing utilities (awk,sed,cut), and bioinformatics tools (bedtools). These commands are used as intended for genomic data processing. Python scripts provided in the skill usesubprocess.runto programmatically invoke these tools, which is a standard pattern for bioinformatics wrappers. - [INDIRECT_PROMPT_INJECTION]: The skill ingests and processes untrusted genomic data files (BED, narrowPeak) provided by the user. While processing external data is an attack surface, the tools involved parse specific structured genomic coordinates rather than natural language instructions, and there is no evidence of dangerous autonomy or lack of sanitization that could be exploited via malicious input data.
Audit Metadata