add-vault-abi-methods
Fail
Audited by Gen Agent Trust Hub on Feb 17, 2026
Risk Level: HIGHCREDENTIALS_UNSAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [CREDENTIALS_UNSAFE] (HIGH): The skill command 'source .local-test.env' (Step 7) loads potentially sensitive environment variables into the process. This file is a standard location for secrets such as private keys and provider URLs, which could be exposed via agent logs or if the agent environment is compromised.- [COMMAND_EXECUTION] (MEDIUM): The skill uses 'poetry run pytest' to execute generated test cases. While a standard development task, the execution of agent-generated code presents a risk if the generation logic is manipulated.- [PROMPT_INJECTION] (LOW): The skill is vulnerable to indirect prompt injection (Category 8) through its ingestion of external ABI files and protocol documentation.
- Ingestion points: Smart contract ABI files in 'eth_defi/abi' and protocol documentation are used to determine code logic.
- Boundary markers: None specified; there are no instructions to sanitize or validate the content of the ABI files against malicious payloads.
- Capability inventory: The agent can write files (Step 3, 5, 6) and execute shell commands (Step 7).
- Sanitization: None present; the agent parses complex JSON ABIs directly into executable Python code.
Recommendations
- AI detected serious security threats
Audit Metadata