idaibin
UserIndependent Agent Skills for repository analysis, planning, implementation, review, delivery, design, and runtime verification.
Categories
Indexed Skills (17)
ui-spec
Use when a selected visual source or accepted UI surface must become a traceable current Feature Spec or Google DESIGN.md contract at a resolved shared design root, including evidence lifecycle, source/runtime deltas, states, responsive/accessibility behavior, assets, and acceptance; route unresolved product behavior to product-spec, current-state audits to audit-frontend, and source edits to dev-frontend.
ask-chatgpt
Use when the user requests a local ChatGPT request package or explicitly authorizes an independent ChatGPT review, research result, visual exploration, or decision challenge; do not use when Codex or an available host tool can complete the result directly.
dev-frontend
Use when a frontend change must be implemented or refactored across UI, state, data, styling, build/tooling, accessibility, performance, or desktop integration, including mapped two-pass visual closure for an accepted selected source; owns source edits and validation, not audit-only, browser-only, UI-spec, or Git delivery.
dev-rust
Use when a Rust source change must be implemented, ported, or refactored across APIs, crates, services, CLIs, async, persistence, unsafe or FFI boundaries, tests, or docs; owns source edits and validation, not audit-only, review-only, or Git-delivery work.
human-writing
Use when drafting, rewriting, proofreading, diagnosing, or adapting source-grounded prose, especially technical articles, essays, tutorials, product notes, and social posts, while preserving facts, attribution, uncertainty, voice, and meaning. Supports private English-first reasoning for Chinese finals; not translation-only, fiction, or AI-detection evasion.
ops-browser
Use when directly operating or verifying a specified page, capturing same-state selected-source/runtime visual and computed evidence, or gathering evidence for an isolated browser-layer failure; require an available browser, verified target, and proven capability, not external-AI orchestration, desktop-client proof, or cross-system diagnosis.
ops-client
Use when directly operating or verifying a specified real desktop client, or gathering evidence for an isolated client-layer failure; require a verified target, window, and capability, not browser-page verification, source changes, or cross-system diagnosis.
product-spec
Use when an ambiguous product feature, new product boundary, named product fact source, or authorized product-document rebuild must become the smallest current implementation-ready specification; owns product decisions and terminal product artifacts, not shared domain-language/lifecycle conflicts, selected-source UI specification, or implementation.
repo-delivery
Use when authorized Git mutation must preserve a large task through semantic milestone, targeted fixup, or exceptional checkpoint commits, normalize task-branch history, or deliver reviewed changes—including a validated Skills catalog release—through commit, push, integration, cleanup, or ref proof; stops before pull-request creation.
repo-map
Use when a Git or non-Git workspace needs a current durable map or authorized rebuild of roots, source/build/runtime identities, ownership, architecture, commands, dependency routes, reusable contracts, Java build boundaries, or layered AGENTS.md guidance; not for task-local discovery or change review.
repo-review
Use when Worktree changes or a fixed snapshot/range need coordinated read-only Standards and Spec findings, documentation-authority review, completed provider-evidence integration, or selected-source visual-completion review; use audit-* for bounded domain audits and a host security workflow for security-only review.
domain-modeling
Use when shared business terms, meanings, rules, lifecycle conflicts, or domain boundaries are ambiguous across product work and must be resolved before specification or implementation; route feature-local behavior and acceptance to product-spec.
ask-ai
Use when the user requests a package or named external-AI result for review, research, cross-review/互审, final-result retention, image work, or an exact saved user instruction such as 进行三方会审; also handles legacy ask-chatgpt wording, but do not use when Codex or an available host tool can complete the result directly.
dev-java
Use when Java source or Java-owned Maven/Gradle configuration must be implemented, migrated, or refactored across Spring services, HTTP/security boundaries, persistence, transactions, messaging, scheduling, caches, tests, or configuration; owns source edits and validation, not non-Java JVM work, audit-only, fixed-basis review, or Git delivery.
ui-design
Use when a concrete page or flow needs evidence-grounded visual and interaction design before frontend implementation, or shared tokens, component semantics, variants, or the product-wide visual system must change.
knowledge-distillation
Distill books, courses, transcripts, SOPs, research, and experience into traceable Knowledge IR and reusable artifacts. Use when Codex must extract principles, decision frameworks, boundaries, anti-patterns, cards, or an Agent Skill from substantial source material while preserving evidence and avoiding unsupported claims.
fixed-basis-repo-review
Review a repository, commit, branch, or fixed diff for actionable defects with explicit evidence and verification limits. Use for review-only requests with an identifiable code basis.
Bio shown is the top-scored skill's repo description as a fallback — real GitHub bios land in a future update.