mise
Pass
Audited by Gen Agent Trust Hub on Jul 23, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill refers to the official installation method for the 'mise' tool via 'https://mise.run | sh'. This is the documented and standard installation path for this well-known developer utility.
- [COMMAND_EXECUTION]: The skill facilitates the execution of user-defined tasks and CLI commands via the 'mise' runtime. It includes a local Bun-based utility script 'Tools/lint.ts' that executes in the user's environment to validate project-level configuration files.
- [DYNAMIC_EXECUTION]: The 'Tools/lint.ts' script employs dynamic imports to load and parse 'mise.toml' files. This is a standard method for the Bun runtime to leverage its built-in TOML and module loading capabilities for configuration parsing.
- [INDIRECT_PROMPT_INJECTION]: The skill processes external configuration data ('mise.toml', 'package.json') which serves as an ingestion surface for untrusted instructions. It mitigates this via the tool's built-in 'trust' mechanism and provides a linter to detect suspicious content.
- Ingestion points: 'mise.toml', '.tool-versions', 'package.json', and project-level task scripts.
- Boundary markers: The skill relies on 'mise trust' which gates the execution of configuration in untrusted directories.
- Capability inventory: Uses 'Bash', 'Write', 'Edit', and 'Grep' tools to manage files and execute shell tasks.
- Sanitization: The 'Tools/lint.ts' script provides automated checks for inlined secrets and house-style consistency.
Audit Metadata