LoopX · public-safe case surface

Showcase & Good Case

Real LoopX cases showing how long-running agent work stays reviewable, verifiable, and safe to continue.

LoopX preserves goals, gates, todos, claims, quota, run history, and evidence outside any single agent session.

LoopX Showcase & Good Case catalog public surface goal / trigger todo / gate / evidence public outcome
01

Top showcase cases

2026-07 · owner_approved_user_evidence_case

13+ hour C++ algorithm accuracy run

An independent user reported a >13h multi-stage C++ accuracy task that stayed on goal, found a public code-memory tool, and improved the final precision.

Proof: A >13h engineering run can remain aligned, change method through public research, and retain an understandable evidence trail.

>13h · user-reported runtime1 · public tool reference0 · parameter-micromanagement reports
Independent userUser report + public method13h plus runvision replan
2026-07 · owner_approved_user_evidence_case

Four-day unattended agent run

An independent user reported that an agent ran for four days without human intervention, continued useful work, and exposed a weekly-report capability.

Proof: A user observed four days of unattended work and judged the work useful, with a later reporting surface available for inspection.

4d · user-reported runtime0 · reported human interventions1 · periodic report surface
Independent userUser reportfour day rununattended execution
2026-07 · public_repository_plus_user_report_case

Public engine refactor across seven merged PRs

A user attributed a public Engine decomposition to LoopX; the repository shows the work landing incrementally through seven merged PRs.

Proof: One public refactor goal landed as seven merged PRs instead of a monolithic change.

7 · public merged PRs1 · public architecture goal1B+ · user-reported token scale
Independent userPublic repo + user reportpublic repositoryseven prs
2026-06-27 · public_evidence_case

Overnight PR batch with reviewable control

An overnight LoopX run can produce many PR-sized slices while keeping review, validation, and public evidence boundaries visible.

Proof: High-throughput multi-lane work can remain PR-sized, reviewable, and merge-safe.

22 · reviewable merged commits10 · PR-numbered commits10h · public evidence window
Creator dogfoodingPublic Git evidencehigh throughput reviewable workpr sized slices
2026-06-24 · public_safe_pattern_case

PR issue automatic fix loop

Review feedback should become an ordered repair workflow with repro, fix, validation, and reviewer handoff.

Proof: Issue and review feedback can enter an executable repair loop.

5 · repair-loop stages1 · branch-ready fix packet1 · review handoff preserved
Reproducible demoPublic-safe patternissue fix workflowreview feedback
2026-06-23 · public_safe_pattern_case

Agent-to-agent PR comment and fix loop

PR review feedback can become an owned agent todo with fix evidence instead of a loose chat reminder.

Proof: Multiple agents can coordinate around PR review comments without losing owner review.

1 · owner per feedback item3 · owner/fix/review questions answered1 · review-packet handoff
Reproducible demoPublic-safe patternagent to agent handoffpr comment loop
2026-06-23 · public_safe_pattern_case

Overnight project refactor as PR-sized slices

A broad refactor can run overnight while staying split into human-sized review units.

Proof: Long unattended refactors can split into moderate PR slices instead of one unreviewable diff.

1 · PR-sized slice at a time3 · successor/supersede/handoff routes1 · review gate for broad risk
Creator dogfoodingPublic-safe patternlong unattended goalpr sized slices
2026-06-19 · public_safe_interactive_case

Dynamic workflow for hardware-agent development

A fuzzy long-running engineering goal needs a shared control plane when multiple worker agents participate.

Proof: Fuzzy goals, multiple workers, and long unattended runs can still converge.

5 · public-safe hardware workflows3 · LoopX/orchestrator/worker role split1 · canonical artifact preserved
Contributor casePublic interactive casedynamic workflowmulti agent coordination
2026-06-19 · public_evidence_case

LoopX self-iteration loop

A high-churn LoopX repo stayed legible while benchmark, product, docs, planning, and peer lanes moved in parallel.

Proof: A high-churn multi-lane project can keep state, boundaries, and evidence coherent.

801 · public commits244 · commits since Jun 1859-92d · AI-assisted baseline
Creator dogfoodingPublic Git evidenceself iterationpeer claim scope
2026-06-17 · reproducible_synthetic_demo

Blocked P0 with safe P1/P2 rotation

A gated P0 lane should not stall a whole long-running goal when safe fallback work exists.

Proof: A user decision should not block all safe work.

1 · concrete P0 user decision1 · safe fallback lane0 · gated-lane auto-progress
Reproducible demoReproducible synthetic evidenceblocked priority fallbackconcrete user gate

Experimental today-value path

A lower-priority entry point for users who want to pick one useful LoopX capability today without replacing the showcase first screen.

PR review/comment -> fix loop Branch-ready fix packet with repro, smoke result, and remaining review owner. Fewer dropped review threads.
Overnight PR-sized refactor Reviewable slice list, validation notes, successor todo, and merge boundary. More merged commits without a giant diff audit.
P0 blocked -> safe fallback Kernel projection of the exact user gate, safe fallback todo, quota decision, and evidence boundary. Less idle agent time while preserving human judgment.
02

Appendix case