Datasets:

cy0307
/

awesome-loop-engineering

Surface	Status	Evidence / next action
GitHub release	Complete	`v0.9.0` packages 730 resources, 22 patterns and contracts, 8 runtime starters, and explicit model-to-operations scope facets.
Release announcement	Ready	Publish the v0.9.0 summary from `posts/launch.md`; Discussion #9 is the low-noise release and contributor feed.
GitHub Explore topic	Complete	`github/explore#5209` merged the `loop-engineering` topic page.
Canonical Awesome directory	Policy objection	`sindresorhus/awesome#4339` is open and lint passes, but the owner stated that AI-generated lists are not accepted. Do not treat further visual polish as a path around that policy or misrepresent automated assistance.
Adjacent harness directory	In review	`ai-boost/awesome-harness-engineering#131` transparently proposes the project for the list's Related Awesome Lists section.
Hugging Face dataset	Complete	`datasets/cy0307/awesome-loop-engineering` carries the structured resource sheet and full project mirror.
English and Chinese launch articles	Ready	Publish `posts/launch.md` and `posts/launch.zh-CN.md` natively on suitable channels.
Profile pin and social posting	Owner action	Pin the repository, then publish from identity-bound LinkedIn, X, Hacker News, and Chinese-language accounts.

Indexing Basics

Confirm the repository is public.
Confirm GitHub Pages is enabled and points to docs/ on main.
Confirm the Pages homepage is set as the repository website.
Confirm docs/sitemap.xml and docs/robots.txt are reachable.
Share the GitHub repository URL and Pages URL from public pages that search engines can crawl.

Search engines may still take days or longer to index a new repository. Backlinks and real activity help discovery.

Hugging Face Dataset Mirror

The Hugging Face dataset datasets/cy0307/awesome-loop-engineering exposes the collection to AI/ML users and tabular workflows. Sync it after each GitHub release or data change.

The mirror tracks the full GitHub tree (docs, patterns, examples, schema, scripts).
The dataset mirror includes generated resource sheets at data/resources.csv and data/resources.jsonl, refreshed from the full English README.md by scripts/export_resource_dataset.py.
The HF copy of README.md is a focused dataset card generated from meta/hf_card_header.yaml and meta/hf_card_body.md by scripts/build_hf_card.py. It documents intended uses, evidence fields, limitations, loading examples, and the latest link status without duplicating the 730-row GitHub README.
The YAML header is HF-only: it must never be added to the GitHub README.md, because the metadata list items break awesome-lint.
Sync uses python3 scripts/build_hf_card.py --output <staging>/README.md followed by hf upload --type dataset against a staging copy. The token lives in the local Hugging Face cache; no token is committed.
Upload from a clean checkout, and exclude local artifacts. hf upload mirrors the working directory as it is on disk, so a sync run from a working copy that holds untracked build or QA output publishes that output too, and later syncs never remove it because the CLI does not delete. A July 2026 sync pushed 79 browser-automation files (.playwright-cli/ console logs and element captures, output/playwright/ screenshots) that were never part of the GitHub tree; they were removed with hf repos delete-files. Sync from a fresh clone, and pass --exclude ".playwright-cli/*" --exclude "output/*" alongside the existing exclusions.
Always pass --exclude "README.md" to the full-tree upload. The GitHub README.md carries no YAML header, so uploading it to the dataset repository overwrites the card and removes the configs block. The dataset is then briefly configured with no explicit data file while data/resources.csv, data/resources.jsonl, and data/resources.parquet all sit in data/, which queues a dataset-viewer job that cannot resolve a split and leaves the viewer stuck on "should be available soon" even after the card is restored. Upload the generated card and data/resources.parquet in the same pass so the repository never holds a card and a shard that disagree.

GitHub-Native Promotion

Use GitHub-native discovery in this order.

Topic page (github/explore). The loop-engineering topic now has a curated page through merged PR github/explore#5209. Keep the topic description ecosystem-wide rather than repository-specific.
sindresorhus/awesome submission. The open PR at sindresorhus/awesome#4339 has a passing lint check and an explicit owner policy objection to AI-generated lists. Keep the curation process transparent and wait for maintainer clarification; do not present the PR as ordinary pending review. See AWESOME_SUBMISSION.md.

Editorial submissions to adjacent lists this repo already cites (harness, context, and agent-paper lists). Submit only where the maintainers explicitly accept related resources. Ready-to-paste entry:

- [Awesome Loop Engineering](https://github.com/ChaoYue0307/awesome-loop-engineering#readme) - Practical resources, reusable patterns, adaptable contracts, and runtime starters for recurring AI-agent systems.

Pin the repository on your profile, and add it to your profile README (profile-level promotion).
About description. Keep this concise value statement aligned with the README and project site: 🔁 Build reliable recurring AI-agent systems: 730 resources, 22 operational patterns, 22 loop contracts, 8 runtime starters, an interactive atlas, and a structured dataset.
Custom social preview (repo Settings, Social preview) so shared links lead with the same four durable proof points. This is UI-only; it cannot be set through the API.
Releases and Discussions provide a lower-noise subscription path. Recommend Watch -> Custom -> Releases and Discussions; do not ask readers to watch every repository event.

Do not use reciprocal-star requests, broad promotional pull requests, or unrelated issue comments. Every directory or newsletter submission should satisfy its normal editorial policy on its own merits.

Measurement Baseline

GitHub retains repository traffic details for only 14 days. Record a snapshot weekly, using the same definitions each time.

Snapshot captured on 2026-07-20 for the API window ending 2026-07-19:

Metric	Baseline
Repository views	637
Unique visitors	157
Clones	720
Unique cloners	199
Stars	35
Forks	7
Watchers	1

The repository overview reached 126 unique visitors. The Chinese README remained the second-most visited content path with 20 unique visitors. The largest identifiable referrers were GitHub and Google (22 unique visitors each), Bing (11), and the project site (8); X contributed two identifiable visitors. Clone traffic remains unusually high and may include automated clients, so do not treat clone counts as human adoption without corroborating forks or contributions.

Since the previous snapshot, stars increased from 31 to 35 and forks from 6 to 7 while watchers remained at 1. Qualified discovery is improving, but release and discussion subscriptions remain the clearest gap; the next campaign should keep prioritizing useful external reach and the low-noise release/digest path.

Working 30-day campaign targets:

1,000-1,500 unique visitors;
75-100 total stars;
8-12 forks;
5-8 watchers;
at least 5 external contributors.

Treat these as measurement targets, not promises. Log the publication date and canonical URL for each channel, then compare referral traffic and qualified contributions rather than optimizing for raw impressions.

Launch Copy

Use the English launch article or Chinese launch article for long-form distribution.

Short Post

🔁 Awesome Loop Engineering maps how recurring AI agents should act, verify, remember, retry, and stop.

730 resources · 22 patterns · 22 contracts · 8 starters

Explore and build:
https://github.com/ChaoYue0307/awesome-loop-engineering

#AIAgents #AgentEngineering

Attach assets/social-preview.png directly when the platform permits it. The post is 264 characters before platform-specific URL shortening.

For a fresh link card on X or LinkedIn, share https://chaoyue0307.github.io/awesome-loop-engineering/x-v10-730.html. The versioned URL points to the current preview image, records the v0.10.0 campaign, and redirects readers to the interactive site.

Chinese Short Post

🔁 Awesome Loop Engineering 把反复运行的 AI 智能体变成可设计、可验证、可停止的工程系统。

669 个资源 · 20 个运行模式 · 20 份循环契约 · 8 个启动模板

浏览并开始构建：
https://github.com/ChaoYue0307/awesome-loop-engineering

#AIAgents #智能体工程

Longer Post

Awesome Loop Engineering v0.9.0 is live.

"Loop" can mean recurrent computation inside a model, reasoning and tools inside one task, or verified work repeated across time. v0.9.0 maps all three without confusing their guarantees.

The release includes:

- 730 papers, docs, tools, benchmarks, and guides
- 29 model-layer resources, including Loopie, labeled as adjacent foundations
- an interactive Resource Atlas
- 22 operational patterns and 22 adaptable loop contracts
- 8 runtime starters: 3 executables and 5 copy/paste templates
- 50-field CSV, JSONL, and Parquet exports mirrored to Hugging Face
- source and evidence fields, with point-in-time link-check results

Explore, reuse, or correct the map:
https://github.com/ChaoYue0307/awesome-loop-engineering

Hacker News

Use a regular submission, not Show HN.

Title: Awesome Loop Engineering: 730 resources from looped models to agent operations

I have been mapping recurring AI-agent systems: how work enters, agents act, evidence gates results, state survives, and the system retries, escalates, or exits.

The repository connects 730 papers, docs, tools, benchmarks, and guides across model, agent, harness, workflow, operations, and evaluation layers with 22 operational patterns, 22 adaptable contracts, 8 runtime starters, an interactive Resource Atlas, and a structured Hugging Face dataset. Model-level recurrence is labeled as an adjacent foundation rather than a complete operational loop. Corrections to source links, taxonomy, and summaries are especially welcome.

https://github.com/ChaoYue0307/awesome-loop-engineering

Do not ask anyone to upvote the submission.

Coordinated Release Sequence

Publish the GitHub release and confirm Pages, dataset, and CI are current.
Publish the English article and a concise LinkedIn or X post using the stack or lifecycle visual.
Publish the native Chinese article rather than translating a social post mechanically.
Submit the regular Hacker News post after the release page is stable.
Contact cited authors with accuracy requests, then approach adjacent list maintainers and newsletters.
Record each channel and its referral result in the 14-day traffic snapshot.

Where To Share

X/Twitter thread with the cover image and working definition.
LinkedIn post focused on the prompt/context/harness/loop stack.
Hacker News or Reddit only if framed as a useful field guide, not a launch announcement.
Relevant GitHub issues or discussions where Loop Engineering, coding agents, or agent workflows are already being discussed.
Authors and maintainers cited in the repo, asking for corrections or better original source links.
The repository's own Discussions, using DISCUSSIONS.md for starter prompts and scope guidance.

Outreach Targets

Use OUTREACH.md for wording.

Direct Loop Engineering article authors.
Agent runtime maintainers with scheduling, goals, hooks, worktrees, skills, plugins, or subagents.
Benchmark maintainers for long-horizon coding agents.
Builders who can contribute real PR babysitting, CI repair, docs drift, deploy verification, benchmark optimization, knowledge freshness, or feedback clustering loops.

Awesome List Submission

Use AWESOME_SUBMISSION.md for the upstream checklist, timing rule, suggested entry, and PR body.

Before submitting to any awesome list:

Verify all links pass.
Confirm the README has a narrow scope and contribution guidelines.
Confirm the license is visible and correct.
Confirm the repo has a release and working CI.
Submit only where the maintainers accept related awesome-list entries.

Candidate places:

sindresorhus/awesome
adjacent prompt, context, harness, and agent engineering lists that accept PRs for related resources

Accuracy Rules

Do not claim that Loop Engineering is a finished standard.
Do not imply cited authors endorse this repository unless they explicitly do.
Describe what readers can explore, compare, or build instead of claiming authority or completeness.
Keep all claims linked to public, inspectable sources.