dev-qa
Warn
Audited by Gen Agent Trust Hub on Sep 9, 2026
Risk Level: MEDIUMDYNAMIC_EXECUTIONINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [DYNAMIC_EXECUTION]: The skill is explicitly designed to rerun documented reproduction or acceptance cases stored in the
.konteksto/audit-register.mdfile. This involves executing code at runtime using tools likepython3ornodebased on data extracted from a markdown file. - [INDIRECT_PROMPT_INJECTION]: The skill ingests data from an external register to drive its logic. If an attacker can modify the audit register to include malicious commands under the guise of a reproduction case, the skill will execute them as part of the QA process.
- [COMMAND_EXECUTION]: The skill utilizes powerful tools including
git,python3,node, and shell utilities (sed,awk) to perform its tasks. The instructions direct the agent to use these tools to run arbitrary logic defined in external documentation.
Audit Metadata