استكشف منشورات X المراجَعة عن المهارات، ومن يشاركونها، والمشاريع المذكورة فيها.
يشمل العدد فقط منشورات المهارات التي جُمعت وروجعت. اليوم الفارغ يعني عدم جمع منشورات هنا. التواريخ بحسب UTC.
275 منشورات جُمعت
Eddi Sanchez opens the HeyEddi CI Skills pack for coding agents. The pack covers eddi-ci.yaml configuration, review findings, failing checks and writing pipeline YAML for isolated Spot runners. The runner pipeline in the source is marked as a placeholder, not a verified execution path.
Hev Mind announces that Layer's public repository provides plain SKILL.md files for agent workflows. The linked documentation covers searching Layer data, consulting its docs, and operating Layer through its CLI.
Scopuly shows how its scopuly-wallet skill can review a Stellar wallet integration for stale account or network state, missing listener cleanup, duplicate signing requests, and unchecked errors. The post distinguishes code findings from behavior requiring real-wallet tests.
Big Iron Chris introduces SageCut for quality review of Remotion videos. He says it uses Levanto's Sage model to inspect frames and sends selected problem frames to Claude; his cost and time figures describe his own tests.
Jakub Kasprzyk introduces the react-native-better-pdf agent skill for migrating from react-native-pdf, covering dependencies, prop mapping, page numbering changes, tests, and the package API.
Wenzi releases two independent podcast skills: one turns an episode link into a readable Simplified Chinese Markdown transcript; the other uses a confirmed creator profile to select viewpoints and draft personal social content from a transcript.
The creator introduces assoomer, a skill intended to keep humans responsible for unclear creative or coding decisions instead of letting an agent guess what a vague request means.
Jeff Li releases chatgpt-fanout-queries for GEO research. It records search queries and sources actually exposed by the OpenAI Responses API, separates retrieved sources from final citations, and creates an offline report; it does not reveal hidden ChatGPT web searches.
Shubhang Rajput introduces design-superskills, a six-skill design workflow covering reference capture, research synthesis, first drafts, Figma organization, design annotations, and developer handoff while keeping creative decisions with the designer.
虹安 describes building vibe-wrapped to read local Claude Code and Codex conversation logs, calculate usage patterns and generate a shareable coding-persona poster. The article says analysis runs locally and notes fixes discovered while testing real logs.
Zain Ali introduces half-fixed, a local git-history scanner and Agent Skill that finds fixes applied to some sibling implementations but missed in others. The agent checks findings before changing code.
Corey Haines shares a content strategy Skill that maps attainable search topics, selects a commercial cluster and produces a hub-and-spoke plan with eight briefs in publishing order and planned internal links.
Leon Lin introduces TasteSkill v2, rebuilt to give coding agents an art direction, source assets and a browser check for their frontend work.
Jared Rhizor introduces the Elmo team's AEO Skill. He says the team compared existing Skills and research, built a larger neutral evaluation set and iterated on the Skill; the post does not establish independent performance results.
Recommends apk-reverse as an Android analysis workflow covering decompilation, DEX changes, repacking and runtime debugging. The repository defines an authorized-testing scope and distinguishes available routes from unverified ones.
Recommends plain-writing for clearer agent prose and explains its command for rewriting the previous answer. The post includes a Chinese rule summary and a before/after example drawn from the project’s evaluation materials.
Highlights motionmaxxing’s motion rules and review scripts through the project’s Cluely comparison, made with the same model and brief. The post discusses moving entrances, easing and cuts during motion.
Recommends penciled for pencil-style architecture diagrams, highlighting its layout checks and the author’s browser-based visual feedback loop. The post explains how a deliberately sketch-like style can invite discussion.
Explains dzhng’s feature workflow: explore unknowns, write a specification, run an implementation loop and review the agent’s recorded decisions. The multi-day run cited in the post is the project author’s example.
Recommends Cloudflare’s security-audit Skill, highlighting architecture mapping, parallel investigation and an independent agent that tries to disprove candidate vulnerabilities.