walrus-cli
Fail
Audited by Gen Agent Trust Hub on Jul 6, 2026
Risk Level: HIGHREMOTE_CODE_EXECUTIONEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [REMOTE_CODE_EXECUTION]: The skill instructs the user to install the suiup tool by piping a script from the developer's official GitHub repository (Mystenlabs/suiup) directly to the shell.
- [EXTERNAL_DOWNLOADS]: Downloads the default client_config.yaml configuration file from the official project documentation domain (docs.wal.app) using curl.
- [COMMAND_EXECUTION]: Executes various subcommands of the walrus CLI to interact with the storage network, including management of file blobs and network status.
- [PROMPT_INJECTION]: Provides a capability to read content from a decentralized network (walrus read), which represents an indirect prompt injection surface.
- Ingestion points: Arbitrary data ingested into the agent context via walrus read in SKILL.md.
- Boundary markers: Absent; no instructions are provided to the agent to isolate or treat retrieved data as untrusted.
- Capability inventory: The skill has the ability to execute subprocesses and write to the local filesystem via the walrus CLI.
- Sanitization: Absent; the skill does not define validation or sanitization for content retrieved from the network.
Recommendations
- HIGH: Downloads and executes remote code from: https://raw.githubusercontent.com/Mystenlabs/suiup/main/install.sh - DO NOT USE without thorough review
Audit Metadata