The Agent Skills Directory

[PROMPT_INJECTION]: The skill processes UI state and screenshots from external web pages during test creation and debugging, creating a surface for indirect prompt injection where malicious content on a site could attempt to influence the agent's behavior.
Ingestion points: momentic_get_session_state tool and artifact files in the .momentic-mcp directory containing serialized UI snapshots.
Boundary markers: The instructions do not define explicit boundary markers or instructions to ignore embedded commands within the ingested UI data.
Capability inventory: The agent has filesystem write access via momentic_test_splice_steps and limited code execution capabilities through test step previews.
Sanitization: No specific sanitization or filtering of external UI content is mentioned before it is processed by the agent.
[REMOTE_CODE_EXECUTION]: The skill supports the use of JAVASCRIPT steps to execute code snippets within the controlled testing environment to handle complex behaviors not covered by native steps.
[COMMAND_EXECUTION]: The skill utilizes a CLI-style syntax for constructing and executing commands that interface with the Momentic MCP tools to manage test sessions and persist changes.

momentic-test