canai-yolo
Warn
Audited by Gen Agent Trust Hub on Oct 2, 2026
Risk Level: MEDIUMREMOTE_CODE_EXECUTIONDYNAMIC_EXECUTIONPRIVILEGE_ESCALATIONPERSISTENCEINDIRECT_PROMPT_INJECTION
Full Analysis
- [REMOTE_CODE_EXECUTION]: The skill describes tools such as
wpcanai-create-snippetandwpcanai-set-snippet-statusthat allow an AI agent to write and live-publish PHP code to a remote WordPress server. - [DYNAMIC_EXECUTION]: The skill facilitates a workflow where the agent generates and modifies executable scripts (PHP, CSS, JS) that are syntax-checked and run in the target environment at runtime.
- [PRIVILEGE_ESCALATION]: Modifying site-wide code snippets provides the agent with administrative-level control over the website's logic and behavior.
- [PERSISTENCE]: Snippets created by the agent are stored on the server and executed according to defined hooks, ensuring that modified behaviors persist across sessions.
- [INDIRECT_PROMPT_INJECTION]: The skill establishes an attack surface where untrusted code or instructions could influence the agent during the snippet management process. Ingestion points: The
codeparameter in snippet creation and update tools in SKILL.md. Boundary markers: The documentation does not specify the use of delimiters to isolate snippet code from agent instructions. Capability inventory: The tools support listing, reading, creating, updating, and publishing code snippets across a WordPress site. Sanitization: The documentation states the server performs syntax validation and a test run of PHP code before saving.
Audit Metadata