cardano-cli-plutus-scripts

Pass

Audited by Gen Agent Trust Hub on Oct 16, 2026

Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [COMMAND_EXECUTION]: Provides a wrapper script scripts/cardano-cli.sh that executes the cardano-cli tool.
  • The script attempts to run the command using a local binary or falls back to a Docker container if the binary is missing.
  • Arguments are passed to the underlying command using the "$@" array expansion, which preserves argument boundaries and mitigates some internal shell injection risks.
  • The script mounts the current working directory ($PWD) into the container, granting the tool access to local files for transaction processing.
  • [INDIRECT_PROMPT_INJECTION]: The skill templates are designed to incorporate user-provided data into executable shell commands.
  • Ingestion points: Untrusted user input such as UTXO hashes, blockchain addresses, and Lovelace amounts are used to populate templates in SKILL.md and reference/script-spend-template.md.
  • Boundary markers: Absent; there are no delimiters or instructions provided to the agent to ignore potentially malicious content within the user-supplied transaction data.
  • Capability inventory: Shell execution through the scripts/cardano-cli.sh wrapper, allowing for blockchain queries, transaction building, signing, and submission.
  • Sanitization: Absent; neither the markdown templates nor the bash wrapper perform validation or escaping of the user-supplied strings before execution.
  • [EXTERNAL_DOWNLOADS]: Fetches a Docker container image from a well-known organization's repository to provide the necessary tooling for the skill.
  • The script pulls the ghcr.io/intersectmbo/cardano-node:latest image, which is maintained by the Intersect MBO organization for the Cardano ecosystem.
Audit Metadata
Risk Level
SAFE
Analyzed
Oct 16, 2026, 11:09 PM