oracle
Warn
Audited by Gen Agent Trust Hub on Sep 8, 2026
Risk Level: MEDIUMEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTIONMETADATA_POISONING
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill instructions use
npx -y @steipete/oracleto fetch and execute a third-party package from the NPM registry without user confirmation. - [INDIRECT_PROMPT_INJECTION]: The skill is designed to read project source files (
src/**) and send them to an external AI model, creating a vulnerability to instructions hidden in the code. - [METADATA_POISONING]: The YAML frontmatter declares no
allowed-tools, which contradicts the skill's reliance on shell command execution for its core functionality.
Audit Metadata