The Agent Skills Directory

[REMOTE_CODE_EXECUTION]: The skill's installation process in SKILL.md and references/setup.md uses the pattern curl -sSL http://agents.allium.so/cli/install.sh | sh. This executes unverified remote code. Additionally, the use of unencrypted HTTP instead of HTTPS makes the script vulnerable to man-in-the-middle attacks.
[COMMAND_EXECUTION]: The skill relies on the execution of various allium CLI subcommands to perform blockchain analytics and realtime lookups.
[CREDENTIALS_UNSAFE]: The authentication setup described in references/setup.md involves the agent collecting and passing highly sensitive information, such as blockchain private keys and API keys, to the CLI tool.
[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted external data which could contain malicious instructions.
Ingestion points: SQL query results from allium explorer and live data from allium realtime commands.
Boundary markers: None identified; the instructions do not specify delimiters to separate untrusted data from the agent's instructions.
Capability inventory: The skill allows the execution of multiple CLI tools and shell-based operations across the blockchain ecosystem.
Sanitization: There is no evidence of sanitization or validation of the data retrieved from blockchain APIs before it is returned to the agent context.

allium-data