bio-rnaseq-qc
Pass
Audited by Gen Agent Trust Hub on Sep 11, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes external file inputs in shell scripts without sufficient sanitization, presenting a command injection surface if the agent handles maliciously named files.
- Ingestion points: Filename arguments (e.g., $1, $2) in
examples/check_rrna.shandexamples/rnaseq_qc.sh. - Boundary markers: None present to distinguish between user-supplied filenames and tool command flags.
- Capability inventory: Execution of system commands including
zcat,grep,sortmerna,java, andawkvia shell subprocesses. - Sanitization: The scripts use unquoted variables (e.g.,
$FASTQ,$BAM) when invoking system commands, allowing shell metacharacters in filenames to be interpreted as code by the shell.
Audit Metadata