validate
Pass
Audited by Gen Agent Trust Hub on Oct 29, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTIONCREDENTIALS_UNSAFE
Full Analysis
- [COMMAND_EXECUTION]: The skill constructs shell commands using data retrieved from local configuration files (such as 'manifest.json' or 'mta.yaml') and direct user input. Maliciously crafted data in these files could attempt to perform command injection when the 'afctl' tool is invoked. \n- [CREDENTIALS_UNSAFE]: The skill instructs the agent to handle SAP credentials. While it specifies secure input methods ('read -s') and warns against logging, the credentials for SAP BTP are passed as command-line arguments to the 'afctl' tool, which may expose them in process listings on the host system. \n- [INDIRECT_PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it reads configuration data from local project files and uses that data to drive subsequent actions, such as fetching runtime URLs or executing CLI commands. \n
- Ingestion points: 'mta.yaml' and 'manifest.json' files located in the project directory. \n
- Boundary markers: Absent; there are no specific delimiters or instructions to ignore embedded commands within the ingested files. \n
- Capability inventory: Shell command execution via Bash ('afctl') and browser automation via Playwright. \n
- Sanitization: Absent; the skill does not explicitly describe validation or escaping of the extracted metadata before use.
Audit Metadata