test-driven-development
Fail
Audited by Gen Agent Trust Hub on Feb 16, 2026
Risk Level: HIGHPROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [PROMPT_INJECTION] (HIGH): The skill presents a high-risk indirect prompt injection vulnerability surface. It instructs the agent to ingest untrusted data (feature and bugfix descriptions) and grants it write/execute capabilities. Evidence Chain: 1. Ingestion Point:
SKILL.md(description: 'implementing any feature or bugfix'). 2. Boundary Markers: Absent; the skill does not define delimiters to wrap external content. 3. Capability Inventory: Code generation and local command execution (npm testinreferences/methodology.md). 4. Sanitization: Absent; no validation or filtering of user input is specified. - [COMMAND_EXECUTION] (LOW): The skill explicitly instructs the agent to execute local shell commands (
npm test) to verify code. While standard for developer tools, this functional capability escalates the severity of potential prompt injection attacks.
Recommendations
- AI detected serious security threats
Audit Metadata