author-webpage-content
Pass
Audited by Gen Agent Trust Hub on Oct 28, 2026
Risk Level: SAFECOMMAND_EXECUTION
Full Analysis
- Command Execution Capability: The skill is configured with access to
Bashand instructions to execute a local version check script (node "${PLUGIN_ROOT}/scripts/check-version.js"). This is a standard operational pattern for ensuring environment compatibility before performing file operations. - Input Sanitization and Safety: The authoritative references explicitly mandate the use of HTML escaping for user-provided content. They also implement strict filtering for URL schemes, specifically rejecting potentially dangerous vectors such as
javascript:,data:, andvbscript:across all interactive components like buttons, links, and media embeds. - Data Handling Boundaries: The skill is designed to operate solely on a local file system within a Power Pages project structure. It contains explicit prohibitions against direct network interactions, Dataverse API calls, or unauthorized credential harvesting, ensuring that changes remain local until a user manually performs a deployment via the PAC CLI.
- Structural Preservation: The instructions emphasize preserving existing site markup, Liquid templates, and security-related headers. It avoids broad search-and-replace operations, requiring the agent to identify narrow edit boundaries to prevent accidental corruption or injection into unrelated sections of the page.
- Indirect Prompt Injection Surface: Since the skill processes user-supplied text to generate HTML, it technically possesses an indirect prompt injection surface. However, the skill mitigates this by enforcing strict markup contracts, requiring human review for ambiguous requests, and forbidding the execution of raw user-supplied scripts or unvetted HTML.
Audit Metadata