roblox-game-workflow-control
Pass
Audited by Gen Agent Trust Hub on Oct 20, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill architecture is designed to ingest and process external project data, including workflow records, asset registers, and documentation sections. This ingestion process creates a surface for indirect prompt injection attacks where malicious instructions could be embedded in the project files.
- Ingestion points: The agent reads from
.workflow-control/work-items/*.json,.workflow-control/project.json,docs/asset-register.json, and project Markdown files specified in work item stages. - Boundary markers: The provided instructions and scripts do not implement explicit boundary markers or delimiters to differentiate project data from agent instructions during processing.
- Capability inventory: The skill provides capabilities for local file system read and write operations restricted to the project repository path via the
scripts/workflow-control.mjsutility. It does not possess network access or system-wide privileges. - Sanitization: The skill's Node.js scripts implement robust security controls, including strict JSON schema validation, identifier format enforcement, and comprehensive path validation that prevents traversal and link-based access to files outside the repository root.
Audit Metadata