[UI-71923] Add UIA hint for interaction that remounts its target

[liviusanduuip]

UI-71923

What

Adds a debugging hint to the uipath-rpa skill for a UI-automation failure mode that a coding agent could not diagnose on its own.

Symptom: an interaction faults with UI element is invalid even though the selector still resolves, and neither increasing the delay nor changing the input mode helps.

Cause: the interaction itself remounts the target as a new DOM node (e.g. TypeInto's built-in click-before-typing, or a click that expands a dropdown/popup). The activity resolved the old node; the next action hits the now-detached one. It is a re-resolution problem, not timing.

Fix: split the interactions into separate activities (e.g. Click then TypeInto) so the second re-resolves its target against the new node.

Changes

• references/uia-elements-interaction-guide.md — new section "Debugging a failed interaction with an element" (generic heading; covers the remount mechanism, the diagnostic tell, and the fix).
• references/ui-automation-guide.md — one-line pointer to that section from Common UIA Pitfalls (keeps the pitfall list terse and routes the agent to the fuller guide).

🤖 Generated with Claude Code

[github-actions]

Claude finished @liviusanduuip's task in 1m 51s —— View job

---

PR Review: docs(uipath-rpa): add UIA hint for interaction that remounts its target

• Read review criteria and project conventions
• Read full diff
• Read full changed files for context
• Verify relative links resolve
• Check cross-skill references
• Check for secrets/tokens/paths
• Post review findings

---

Summary

Adds a debugging hint for a UIA failure mode where an interaction remounts its target DOM node, causing a stale-node fault. The content is split correctly: a terse pitfall pointer in ui-automation-guide.md and the full explanation in uia-elements-interaction-guide.md.

Change-by-Change Review

1. skills/uipath-rpa/references/ui-automation-guide.md (+1 line)

Severity: Low

New pitfall entry at line 208 follows the established pattern: one-line symptom + link to the detailed guide. Correctly avoids duplicating the mechanism/fix (Reference Pointer Discipline: ✅). No UIA boundary violations — only public activity names (TypeInto, Click), no forbidden subcommands/flags/artifacts.

One minor nit: the title reads "An interaction faults though the selector still resolves" — "though" works grammatically (as "although"), but "even though" would match the stronger contrast intended and read less like a typo of "through". Very minor.

2. skills/uipath-rpa/references/uia-elements-interaction-guide.md (+8 lines)

Severity: OK

New section ### Debugging a failed interaction with an element (lines 79–85) is well-placed between dropdowns/comboboxes and the existing "Buttons disabled during async operations" section. Content is clear and actionable:

• Explains the remount mechanism concisely
• Gives the diagnostic tell (selector resolves, delay/input-mode don't help)
• States the fix (split into separate activities)
• No UIA boundary violations
• Prose follows token-optimization conventions (terse, prescriptive)

The section heading is generic ("Debugging a failed interaction with an element") — as noted in the PR description, this is intentional to allow future expansion. Works well.

What's Missing

Nothing identified. This is a focused two-file documentation addition with correct cross-referencing.

Area Ratings

Area	Status	Notes
Frontmatter	OK	No frontmatter changes
E2E Tests	OK	Documentation-only change, no new CLI workflows or commands
Skill Body	OK	SKILL.md not modified
References & Assets	OK	New section correctly placed; link resolves; pointer discipline followed
Repo Hygiene	OK	No secrets, no cross-skill refs, changes scoped to uipath-rpa

Issues for Manual Review

• Domain accuracy: The remount-on-focus / click-before-typing failure mode described is based on real UIA behavior — but the reviewer should confirm that TypeInto's built-in click-before-typing is indeed a documented compound action that resolves the target only once (not twice). This is a domain-specific claim the automated review cannot verify.

Conclusion

Clean PR. The content is well-structured, follows the reference pointer discipline, respects the UIA boundary rules, and adds genuine diagnostic value for a failure mode that agents struggle with. The only finding is a cosmetic "though" → "even though" wording nit (Low severity). Approve.

---

[alexandru-petrescu]

/cherrypick release/v1.197 UI-71923

[alexandru-petrescu]

/cherrypick release/v1.198 UI-71923

[github-actions]

A cherry pick from this PR targeting release/v1.197 was triggered.

[github-actions]

A cherry pick from this PR targeting release/v1.198 was triggered.

[github-actions]

⚠️ Cherry-pick to release/v1.197 for UI-71923 opened with conflicts: #2106 — resolve them on the branch before merging.

[github-actions]

⚠️ Cherry-pick to release/v1.198 for UI-71923 opened with conflicts: #2107 — resolve them on the branch before merging.