build-mule-integration
Pass
Audited by Gen Agent Trust Hub on Sep 18, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill follows secure development practices, including a two-phase architecture that separates design from execution and requires a hard user-approval gate (Step 7) before any project files are created or modified.
- [EXTERNAL_DOWNLOADS]: The skill references and assists in the installation of official, trusted vendor tools from the npm registry, such as
@mulesoft/anypoint-cli-v4and@salesforce/anypoint-cli-dx-mule-plugin. These are well-known resources provided by the skill author (mulesoft/salesforce). - [COMMAND_EXECUTION]: The skill utilizes a set of Bash scripts to interact with the local environment and the Anypoint CLI. These scripts are used for legitimate technical tasks like validating prerequisites, searching the Anypoint Exchange, and linting Mule XML files. The execution is scoped to the developer's project context.
- [DATA_EXPOSURE]: The skill interacts with the standard configuration files of the Anypoint CLI (e.g.,
~/.mule-dx/config.json) to resolve environment paths and sessions. This is necessary functionality for the tool's integration with the MuleSoft ecosystem and does not involve exfiltration to untrusted domains. - [INDIRECT_PROMPT_INJECTION]: The skill processes external metadata from Anypoint Exchange. It mitigates potential injection risks by using structured data parsing with
jq, performing static validation of generated XML before building, and presenting all findings in a technical summary for human verification. - [DYNAMIC_EXECUTION]: Benign dynamic execution is used in
scripts/validate_before_build.sh, which calls a Python helper (_suggest_nearest.py) to provide fuzzy-match suggestions for error types. This is a local utility function for development validation.
Audit Metadata