env-to-fnox
Pass
Audited by Gen Agent Trust Hub on Oct 9, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill instructs the agent to ingest and process data from local
.envfiles, which constitutes an attack surface for indirect prompt injection if the files contain malicious instructions. - Ingestion points: The agent is instructed to read the contents of
.envfiles in Step 1 using thecatcommand. - Boundary markers: No specific delimiters or warnings to ignore embedded instructions within the
.envcontent are provided. - Capability inventory: The skill utilizes file reading (
cat,rm), version control (git), and secret management tools (op,fnox,mise). - Sanitization: There is no evidence of sanitization or filtering of the content read from the environment files before categorization.
- [COMMAND_EXECUTION]: The skill uses various command-line tools to perform the migration. These are standard developer tools and their use here is benign and functional.
- Evidence: Use of
op(1Password CLI) for secret creation,misefor tool management, andfnoxfor configuration and environment export.
Audit Metadata