dojo-review
Pass
Audited by Gen Agent Trust Hub on Oct 5, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted external code using Read, Grep, and Glob tools, creating an attack surface for instructions embedded in the analyzed code.
- Ingestion points: Project source files (models, systems, tests) accessed via Read, Grep, and Glob in SKILL.md.
- Boundary markers: Absent; the skill does not define specific delimiters for the content it reads.
- Capability inventory: Limited to Read, Grep, and Glob. No network, file-write, or shell execution capabilities are requested or used.
- Sanitization: Absent; the skill relies on the agent's internal logic for processing read content.
- [SAFE]: The skill provides constructive security guidance, including examples of authorization checks and integer overflow prevention in smart contracts. The external resources mentioned (related skills) appear to be part of the author's own ecosystem.
Audit Metadata