oma-translation
Pass
Audited by Gen Agent Trust Hub on Oct 11, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes external, untrusted content (source text and locale files), which creates an attack surface for indirect prompt injection where malicious instructions could be embedded in the data being translated.\n
- Ingestion points: As specified in
SKILL.md, the skill ingests 'Source text', 'target language or locale', and 'Existing locale files' as its primary operational data.\n - Boundary markers: The instructions in
SKILL.mddo not explicitly mandate the use of robust structural delimiters (such as XML tags) or specific warnings to the agent to ignore any instructions found within the source content.\n - Capability inventory:
SKILL.mdidentifies 'Native search/read' and the ability to 'Edit files' as core capabilities, which provides a functional path for potentially malicious instructions to affect the local filesystem.\n - Sanitization: The provided documentation and language profiles (en.md, ja.md, ko.md, zh.md) do not describe any sanitization, filtering, or escaping mechanisms for the external content before it is processed by the agent.
Audit Metadata