angry-customer-playbook
Pass
Audited by Gen Agent Trust Hub on Jul 1, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [PROMPT_INJECTION]: The skill incorporates untrusted data from customer messages directly into its prompt context, creating an indirect prompt injection surface.
- Ingestion points: The $ARGUMENTS variable in SKILL.md receives untrusted input.
- Boundary markers: Absent. The user-provided text is not isolated from instructions using delimiters or explicit 'ignore' directives.
- Capability inventory: No scripts or tool capabilities were found; the skill explicitly disables model invocation in the frontmatter.
- Sanitization: No input validation or filtering is performed.
Audit Metadata