hreflang-international
Pass
Audited by Gen Agent Trust Hub on Aug 15, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [PROMPT_INJECTION]: The skill contains a directive to "follow it as the active workflow" pointing to an external file. This is an instruction-redirection pattern that tells the agent to disregard its current context in favor of the content found at the specified path, which is a common technique for overriding safety constraints.
- [INDIRECT_PROMPT_INJECTION]: The skill utilizes directory traversal sequences (
../../) to point the agent toward a file outside its immediate skill folder. This creates an attack surface for reading files on the filesystem that may be outside the intended scope of the skill. - Ingestion points: The skill instructs the agent to read and follow the file at
../../seo/hreflang-international/SKILL.md. - Boundary markers: Absent; there are no instructions to the agent to treat the imported content as untrusted or to ignore embedded instructions within it.
- Capability inventory: The skill requires the ability to resolve relative paths and read file content from the local environment.
- Sanitization: None; the skill does not validate the integrity or the location of the target file before instructing the agent to adopt its logic.
Audit Metadata