Explorez des publications vérifiées sur X, les personnes qui les partagent et les projets qu’elles mentionnent.
Seules les publications sur les Skills recensées et validées sont comptées. Une journée vide signifie qu’aucune publication n’a été recensée ici. Dates en UTC.
65 publications recensées
Uriah, an actual contributor to impeccable-rust, shares the Rust review and verification Skill and says he used it for OpenGrok. Its source organizes checks around failure risks and explicit evidence limits.
animation-performance gathers rules for diagnosing dropped frames, including animated shadows, large blur values, inherited CSS variables, and a busy main thread.
Oil UI’s author reports new insights and simpler debugging scripts. The post’s pictured works use Oil UI Pro; the linked open-source Skill focuses on design directions, comparison, and screenshot-based refinement.
backend-performance-review examines backend code with evidence and workload context, and forbids invented performance numbers. The creator shares a test series comparing speculative review advice with findings that identify their evidence and unknowns.
jev-audit reviews a repository for bounded decisions that may suit Jev models, then presents candidate integrations and their tradeoffs. The creator shares the Skill and invites people to report useful findings.
Bole Chen open-sources serverless-cost-guardrails to self-check serverless setups for paths that can produce surprise runaway bills, and invites trials and PRs.
Bhavesh Chowdhury shares a Skill that helps AI agents build, adapt, or review React and Next.js charts with @kind-ui/charts—covering chart selection, composition, styling, legends, loading, and accessibility.
Jared Palmer quotes a write-up on fine-tuning Kev-4B and points to the official kev-finetune Skill, which finds decision-model callsites, helps generate synthetic data, walks through fine-tuning on Modal, and pairs with kev-deploy.
tinkeredbyerwin presents grip inside coding agents: before push, five questions on the agent-written diff that you must answer; below the pass mark, push is blocked so humans still own AI-written code.
Paul Hudson announces SwiftUI Pro agent Skill v2.0 with improvements from real-world iOS 27 and iPhone Duo testing, plus advice drawn from Apple's own skills.
Zeke updated faster-gh-cli-skill with guidance for stacked pull requests, which split a large change into dependent PRs. The linked GitHub revision contains the new section; the catalog still shows the earlier source.
ten-principles applies Ferris Kleier's adaptation of Dieter Rams' design principles to code naming, comments, structure, tests and redundancy. The creator distinguishes this code-style review from architecture or design-pattern decisions.
ollama-decision-skills is a set of Skills for building local decision-model prototypes, integrating typed decisions and handling model imports. The creator illustrates the idea with file organization and personalized feed filtering.
gdp-rs pairs Rust runtime checks with evidence tied to the exact user and resource of a protected operation. Its Agent Skill guides adoption, lifetime errors and Cargo diagnostics.
Damian Galarza wrote a Skill to guide agents through GitHub stacked PRs, after repeatedly correcting how they used the feature. It covers splitting a large change into focused, dependent reviews.
The gh-stack CLI extension includes an Agent Skill for splitting dependent changes into a stack of branches and pull requests. Each PR uses the branch below it as its base, so reviewers can examine one layer at a time.
Antoine shares three updates to SwiftUI Expert Skill for adapting iPhone Duo layouts across fold states. The linked 5.2.1 release extends guidance drawn from Apple resources, Xcode’s bundled Skill and his own app work. The Skill is available under MIT.
A skill for tuning self-hosted LLM inference around the model, hardware, API requirements, and serving workload, including parallelism, KV tiers, and prefill/decode separation.
Measures a research replication package's peak RAM and memory trace, including child processes, then records the measured requirements and evidence in its README.
Review SwiftUI code for deprecated APIs, accessibility gaps and performance pitfalls. SwiftUI Pro reports issues by file with suggested fixes.