typescript
Pass
Audited by Gen Agent Trust Hub on Oct 8, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface because its logic is driven by data ingested from external, potentially untrusted project files.
- Ingestion points: The agent reads the version field from node_modules/typescript/package.json and searches project lockfiles (SKILL.md).
- Boundary markers: No delimiters or protective instructions are defined to separate the ingested version data from the agent's instruction context.
- Capability inventory: The skill allows the agent to edit application source code (.ts, .tsx) and configuration (tsconfig.json), and mentions executing shell commands like tsc (SKILL.md, references/7.md).
- Sanitization: There are no instructions to validate or sanitize the version strings or file contents retrieved from the project environment.
Audit Metadata