input-validation-checker
Pass
Audited by Gen Agent Trust Hub on Sep 9, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONNO_CODE
Full Analysis
- [SAFE]: The provided file contains only markdown documentation and frontmatter configuration with no evidence of malicious intent, obfuscation, or unauthorized data access. The frontmatter uses 'allowed-tools' to restrict rather than expand the agent's capabilities.
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to process user-supplied security requirements to generate production-ready code and configurations, which creates a potential surface for indirect prompt injection.
- Ingestion points: User-provided requests for security patterns and validation logic mentioned in the 'When to Use' and 'Examples' sections of SKILL.md.
- Boundary markers: Absent; the instructions do not provide explicit delimiters or guidance for the agent to separate user-provided data from instruction logic.
- Capability inventory: The skill permits the use of 'Write' and 'Bash' tools, which allow for the creation and execution of code generated from user input.
- Sanitization: The instruction set contains no procedures for validating, escaping, or sanitizing the input data before it is interpolated into code templates.
- [NO_CODE]: The skill consists entirely of instructional markdown and does not include any accompanying script files, binaries, or external dependencies.
Audit Metadata