Cycle 2 discovery: rank AI and gaming workflow painpoints #7

Purpose

Deliver a decision-ready research and repository-fit assessment for the Cycle 2 work in umbrella issue #5. This is the gate before any candidate implementation is selected. Do not treat the current controller-compatibility hypothesis as a decision.

Research deliverables

• Refresh direct evidence from active AI and gaming communities, including Steam Community, relevant GitHub issues/discussions, AI developer forums, Hacker News, and accessibility-focused sources where reachable. Record unsuccessful searches (including Reddit if still blocked) and meaningful coverage gaps.
• For each source, record its direct URL, post/report date, user type and situation, reported pain, workaround, recurrence/engagement evidence, and what remains unknown. Separate the source's claims from our interpretation; do not estimate prevalence from thread activity.
• Audit current OpenController packages, documentation, open issues, and merged work to identify what the SDK already supports and what it can control. Exclude upstream bugs and perception/game-specific features unless a bounded integration belongs in this repository.
• Rank at least three viable opportunity themes, where evidence allows, by user value, evidence strength, project fit, implementation cost, and platform/security risk. Carry forward low-confidence options only with explicit caveats.
• Recommend one bounded implementation direction, or explain why the available evidence does not justify starting a build. Link the fleet-created candidate issues and update their dependency/selection state.
• Publish a concise source log and decision record in this issue or a repository roadmap update, with a task DAG and measurable acceptance criteria for the selected work.

Existing signals to recheck

The Cycle 1 roadmap recorded examples of device-specific mapping/calibration problems in SDL #15658, SDL #16237, and SDL #16434; Steam Input lifecycle/input-loss reports in Steam for Linux #13665; and agent-evaluation error/replay friction in FLE #417 and FLE PR #413. These are examples, not proof of prevalence. Verify current state and find counterexamples before prioritizing.

Acceptance criteria

• A reviewer can trace every selected opportunity to dated, direct evidence and see the limits of that evidence.
• At least two distinct reports support the selected recurring theme when available; a single report remains explicitly a low-confidence lead.
• The repository fit is demonstrated against current code and docs, with clear product-controlled boundaries.
• The selected build issue(s) have scoped tasks, dependencies, non-goals, and acceptance criteria. Unselected candidate issues remain visibly deferred.
• Research and decision work is complete before agents begin implementation; no candidate issue is assumed approved solely because it exists.

Dependency

This is a child work item of umbrella Cycle 2 issue #5 and gates all implementation candidates created for the cycle. Complete this decision before selecting a build slice.