ci-images
Warn
Audited by Gen Agent Trust Hub on Nov 8, 2026
Risk Level: MEDIUMDYNAMIC_EXECUTIONINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [DYNAMIC_EXECUTION]: The script
scripts/release-readiness.shutilizesbash -lc "$cmd"to execute shell commands provided via command-line arguments such as--generateand--test. This enables the execution of arbitrary shell code assembled at runtime. - [INDIRECT_PROMPT_INJECTION]: The skill is designed to extract and execute CI-related commands from the local repository environment, which could be exploited if an attacker places malicious commands in project files like a
MakefileorREADME.mdthat the agent then passes to the execution script. 1. Ingestion points: CI-equivalent commands (e.g.,maketargets) gathered from the local repository as specified inSKILL.md. 2. Boundary markers: Absent; the instructions do not include delimiters or safety warnings for the agent when processing repo-sourced commands. 3. Capability inventory: Thescripts/release-readiness.shfile provides full shell command execution capabilities via therun_stepfunction. 4. Sanitization: Absent; the script does not validate or sanitize the command strings before execution.
Audit Metadata