teach
Pass
Audited by Gen Agent Trust Hub on Sep 15, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is instructed to ingest and analyze untrusted data provided by users to facilitate teaching. This creates a potential vulnerability where an attacker could embed malicious instructions within the code or diffs being explained to influence agent behavior.
- Ingestion points: The instructions mandate reading "the code yourself to get oriented" and referencing "the diff, the code, or the debugger" (SKILL.md).
- Boundary markers: The skill does not define specific delimiters or instructions to ignore embedded commands within the ingested data.
- Capability inventory: The skill invokes other internal skills (how, why, unslop) and uses image generation and Mermaid diagram tools. It does not perform sensitive file operations, network requests, or direct shell command execution.
- Sanitization: There are no provisions for sanitizing or escaping the content read from the codebase before it is processed by the LLM.
Audit Metadata