zapier-make-patterns
Pass
Audited by Gen Agent Trust Hub on Nov 3, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONMETADATA_POISONING
Full Analysis
- [SAFE]: The skill is entirely instructional, providing design patterns and examples for third-party automation tools. It does not contain any executable scripts, shell commands, or network operations.
- [INDIRECT_PROMPT_INJECTION]: The skill documents workflows that ingest untrusted data from external sources such as Gmail, webhooks, and Typeform.
- Ingestion points: Triggers from Gmail, webhooks, and Typeform modules as described in SKILL.md and references/detailed-guide.md.
- Boundary markers: The guide recommends using filters and synthetic data testing to prevent unauthorized runs.
- Capability inventory: None; the skill is a text-based documentation asset without access to shell or network tools.
- Sanitization: The skill describes data transformation patterns, including text manipulation and lookup tables, to clean and validate inputs.
- [METADATA_POISONING]: The skill's frontmatter and detailed guide include self-assigned risk labels (e.g., critical, high) that refer to the reliability of business processes and automation tasks. These are evaluated as informative context regarding platform behavior rather than security vulnerabilities.
Audit Metadata