dx-scratch-def-configure
Pass
Audited by Gen Agent Trust Hub on Oct 30, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADSCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes natural language descriptions of product requirements to generate technical configuration files, which is an inherent risk of the skill's primary purpose.
- Ingestion points: Processes natural language product asks as defined in SKILL.md.
- Boundary markers: Implements a mandatory "Preview" gate with a confirmation card (confirm-card.md) to ensure the user reviews and consents to the draft before any file is written.
- Capability inventory: Writes configuration files to the config/ directory and generates report.md files on disk.
- Sanitization: Employs deterministic mapping of features against a known catalog, uses regex to sanitize org names, and performs JSON schema validation if the jsonschema library is present.
- [EXTERNAL_DOWNLOADS]: Fetches updated feature definitions and configuration settings from official Salesforce developer documentation (developer.salesforce.com). These downloads target a trusted vendor domain and the content is used strictly as a data catalog for mapping, not for execution.
- [COMMAND_EXECUTION]: Invokes bundled Python scripts (lib_scratch_def.py) to perform validation and enrichment tasks. The command execution is scoped to the skill's own logic and is used to automate project configuration based on deterministic rules.
- [SAFE]: The skill demonstrates best practices by validating project roots, preventing the creation of sensitive directories, and requiring explicit user consent before performing file operations. All external resources are sourced from the vendor's own infrastructure.
Audit Metadata