wp-speculative-loading
Pass
Audited by Gen Agent Trust Hub on Sep 9, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted site data (URLs and custom rule patterns) to generate WordPress configuration code, creating a potential surface for indirect prompt injection.\n
- Ingestion points: Site URLs and path patterns processed in
SKILL.md(e.g., inwp_speculation_rules_href_exclude_pathsandwp_load_speculation_ruleshooks).\n - Boundary markers: The skill does not define specific boundary markers or delimiters for untrusted data input.\n
- Capability inventory: The agent can generate PHP code blocks and modify site-wide speculation rule sets.\n
- Sanitization: The documentation notes that WordPress Core sanitizes inputs, but the example code snippets do not implement explicit validation or escaping for the input data they handle.\n- [SAFE]: The skill implements security best practices by including a 'Safety model' section. This section provides critical guidance on avoiding prefetching for state-changing operations (GET requests that mutate data, log out users, or trigger billing), which is a key security consideration for the Speculation Rules API.
Audit Metadata