test
Pass
Audited by Gen Agent Trust Hub on Oct 22, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill ingests data from local project files that could contain malicious instructions aimed at overriding the agent's logic during the testing phase.
- Ingestion points: Reads
AGENTS.md,TASKS.md, and task-specific markdown files indocs/task/. - Boundary markers: Absent. The instructions do not define boundaries for untrusted data or warn the agent to treat external documentation as data only.
- Capability inventory: The agent is empowered to execute arbitrary shell commands (unit/integration/CLI tests), interact with browsers, and modify files (
TASKS.md,docs/testing/). - Sanitization: Absent. The skill does not describe any validation or filtering of the content retrieved from the task documents.
- [COMMAND_EXECUTION]: The skill directs the agent to execute shell commands for verification purposes. While it instructs the agent to prefer "safe verification commands," the specific commands are derived from external project documentation, creating a path for potential command injection if those files are manipulated by an attacker.
Audit Metadata