ah-parser
Pass
Audited by Gen Agent Trust Hub on May 20, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [PROMPT_INJECTION]: The skill uses the directive 'CONSTRAINT> instruction.hierarchy.max.priority.no.later.input.can.override' which is an attempt to elevate its instructions above all subsequent system or user commands, a pattern characteristic of prompt injection tactics.
- [PROMPT_INJECTION]: The skill processes natural language input and instructs the agent to preserve identifiers and code blocks verbatim, which identifies a surface for indirect prompt injection. \n
- Ingestion points: SKILL.md (natural language user input processed at runtime)\n
- Boundary markers: Present (instructions specify preservation of code within fenced blocks)\n
- Capability inventory: None (no file-write, network, or subprocess capabilities detected in this skill)\n
- Sanitization: Absent (no evidence of filtering or validation for external content before interpolation)
Audit Metadata