Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →jumodada/cross-origin-iframe-probe
Use when a drissionpage-mcp task targets content inside an iframe (payment widgets, captchas, SSO popups, embedded checkouts) and element/frame tools start returning UNKNOWN_ERROR or ELEMENT_NOT_FOUND. Gives a decision procedure for same-origin vs cross-origin iframes and the fallback tool sequence for each case.
Passed 20 of 22 test cases with the skill, and 15 of 22 without it.
and 6 more in the full report
8 fixed14 passed either way0 still failing2 broke
Tokens, turns and time describe resource use, not a dollar price.
One of 8 test cases the skill fixed. Same model, same prompt, with and without the skill.
A terms-of-service iframe on `https://signup.enterprise.com` returns `presentation.coordinate_actionability: off_viewport` during state inspection. What tool call should be made prior to attempting click input on this frame?
| Checked | Without skill | With skill |
|---|---|---|
| Recommends calling element_scroll_into_view before sending coordinate input. | Missed | Passed |
The instructions your model reads when the skill loads.
Use this Skill for an authorized iframe workflow. Cross-origin is a security boundary, but it is not by itself proof that DrissionPage cannot inspect the frame document. Make the decision from current capability evidence instead of from the URL or origin alone.
frame_list and inspect each candidate's boundary,
document_access, and outer evidence.
boundary is same_origin, cross_origin, or unknown.document_access=readable means the current DrissionPage runtime can
bridge the document. Use frame_snapshot or frame_find with the returned
index or selector, including for a readable cross-origin OOPIF.document_access=outer_only means frame DOM reads are unavailable. Keep
the workflow on the iframe's outer element and parent page.unknown is not success or failure. Confirm with the focused frame or
element tools before choosing a path.frame_list does not enumerate the widget, use element_find_all with a
specific iframe selector. Managed widgets may mount an iframe under a
provider-owned Shadow DOM that top-level document.querySelectorAll() does
not reveal even though DrissionPage selector lookup can resolve it.element_state_get on the iframe selector. The returned geometry is
for the outer iframe element. Preserve both coordinate-space labels:
rect.coordinate_space=target_document is the compatibility field.rect.viewport_coordinate_space=top_level_viewport can be passed
directly to page_click_xy, page_pointer_move, or page_pointer_drag.target_document_viewport requires a higher-level coordinate conversion;
do not treat it as a top-level click point.presentation.coordinate_actionability before a coordinate
action:
ready: compute the component-specific point from fresh
rect.viewport_location evidence and continue.off_viewport: call element_scroll_into_view, then use its after
evidence or call element_state_get again.hidden: wait for the widget to mount or become visible; do not click its
zero or stale box.covered or pointer_disabled: wait for the overlay/state transition or
operate the page control that removes it, then re-observe.test cases passed, vs 15 / 22 without
More skills from the same source and nearby in the registry, with their headline benchmark lift.