bootstrap-env-banner
Pass
Audited by Gen Agent Trust Hub on Oct 4, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill audits local project files and accepts user-supplied configuration to generate code changes, creating a vulnerability surface if the repository or input is compromised.- Ingestion points: The skill reads
mix.exs,config/runtime.exs, andfly.toml(SKILL.md Phase 0) to detect application metadata, and processes user input for the production host and banner label (SKILL.md Phase 1).- Boundary markers: Absent; the skill does not use delimiters or warnings to ignore instructions that might be embedded in the audited files.- Capability inventory: File-write access toconfig/config.exs,lib/my_app_web/components/layouts.ex, and all root layout HEEx templates (SKILL.md Phase 2).- Sanitization: No explicit validation, escaping, or filtering is performed on values extracted from the project environment or user input before they are interpolated into the generated Elixir configuration and component code.
Audit Metadata