gigio1023

@gigio1023 source repo

51 skills
0 followers
11 hours ago last updated

▌
Astro Dev · gigio1023 bundle
Use when editing .astro/.mdx files, astro.config.*, Astro content collections, client directives, Actions/forms, Tailwind CSS, view transitions, server features, adapters, or upgrades to Astro 7. Routes to current Astro 7 and Zod 4 patterns while guarding against removed Astro 3/4/5/6 APIs. NOT for generic framework-agnostic frontend design or non-Astro projects.

0
installs
▌
Godot Best Practice · gigio1023 bundle
Use when implementing, debugging, reviewing, or upgrading a Godot project, including project.godot, GDScript, Godot C#, .gd, .tscn, .tres, scenes, resources, signals, nodes, UI, physics, shaders, imports, exports, addons, or engine-version changes. Also trigger on Godot resource terms such as ExtResource, SubResource, NodePath, and res:// even when "Godot" is omitted. NOT for engine-neutral game design, unrelated C#/GLSL work, asset creation alone, or contributing to the Godot engine source itself.

0
installs
▌
Unity Game Dev · gigio1023 bundle
Use when planning, implementing, debugging, or reviewing gameplay in an existing Unity project, including C# under Assets, scenes, prefabs, packages, input, UI, physics, NavMesh, Editor automation, Unity AI/MCP, Sentis, ML-Agents, tests, and performance. Adapts to the project's Unity version and available Editor-control capability. NOT for engine-neutral game design, ordinary non-Unity C#, standalone asset creation, generic ML work, or contributing to the Unity engine source.

0
installs
▌
Korean Clarity · gigio1023 bundle
Use when drafting or revising Korean agent responses or artifacts where compression has omitted or blurred sentence components, particles, endings, predicates, logical relations, referents, or ordinary technical wording; or when the user asks to make Korean meaning clearer without changing facts. Restores semantic completeness without padding or invented context. NOT for translation, grammar- or spelling-only proofreading, AI-authorship detection, document structure, evidence, or voice problems without a Korean clarity defect, or changing code, identifiers, commands, quotations, and established product terms.

0
installs
▌
Slop Aware Writing · gigio1023 bundle
Use for writing or revising documents in English, Korean, Italian, or Chinese when the work must prevent or remove AI slop while preserving meaning, evidence boundaries, reader context, and the writer's voice. Covers humanizing or deslop requests ("AI 티 빼줘"), formulaic or generic model-assisted drafts, session-dependent documents, terminology review, and slop-aware authoring of README, guides, specs, ADRs, memos, and posts. Also handles intra-Chinese locale or script normalization. NOT for grammar- or spelling-only proofreading, translation between languages, Python docstrings, prompt coaching, diagrams, PR/commit/issue copy, or implementation changes.

0
installs
▌
Data Chart · gigio1023 bundle
Use when the user wants a data chart, graph, or plot in the clean editorial blog-figure look: line, bar, scatter, or dot-plot renderings of measured data, produced programmatically with matplotlib and exported to SVG/PNG. Trigger on "chart", "graph", "plot", benchmark-score figures, and requests for minimal blog-style data graphics. NOT for flowcharts, architecture diagrams, or box-and-arrow figures (use technical-diagram), and NOT for interactive dashboards or web-embedded live charts.

0
installs
▌
Drawio Diagram · gigio1023 bundle
Use when the user explicitly wants a native .drawio artifact, asks to edit an existing draw.io file, needs draw.io metadata or pages, or requests draw.io export to SVG, PNG, or PDF. Produces editable mxGraph XML with automatic layout first, manual routing only when needed, and structural plus rendered quality checks. NOT for generic technical diagrams or measured data charts.

0
installs
▌
Technical Diagram · gigio1023 bundle
Use when the user wants a technical diagram, architecture figure, system map, process flow, or box-and-arrow schematic without requiring a native editor format. Produces a compact D2 source with ELK layout and SVG output, using a restrained editorial style and rendered quality checks. NOT for native .drawio files, Mermaid-only output, measured data charts, or image generation.

0
installs
▌
Game Review · gigio1023 bundle
Use when reviewing a stable game-design claim, playable build, prototype, milestone or release readiness, narrative content, accessibility, or playtest results, or judging whether a claim like "is this fun", "ready to ship", "playable", "vertical slice", or "done" is backed by real play and ready for a sign-off. NOT for choosing, repairing, or auditing unresolved direction artifacts (use game-direction) and NOT for planning or implementing work (use game-production).

0
installs
▌
Game Direction · gigio1023 bundle
Use when game work starts without a stable direction: a creator asking "what game should I make", turning taste or references into a concept and creative brief, adopting an existing direction or pitch, judging "is this idea worth building" or whether to continue or pivot or kill an existing project, reviewing whether direction artifacts are stable enough for production, or auditing an ambiguous game repo, prototype, or notes. NOT for milestone planning or implementation (use game-production) and NOT for reviewing builds or readiness claims (use game-review).

0
installs
▌
Game Knowledge · gigio1023 bundle
Use when a game question needs field knowledge rather than a process: "screen shake", "hit-stop", "juice", floaty or flat feel, camera complaints, pricing, wishlist numbers, Next Fest, store page and demo scope, finding playtesters, burnout, IP assignment and contractor agreements, localization, publishers, or the design theory behind a mechanic. Also use on first contact with a game project — project.godot, Assets/, *.uproject, or a package.json with three/phaser/babylon — and when a project is months old with no marketing, no playtesters, or no backups. NOT for defining direction (use game-direction), planning or implementing milestones (use game-production), or judging a build or a readiness claim (use game-review).

0
installs
▌
Game Production · gigio1023 bundle
Use when writing a "milestone plan" or scoping prototypes, a "vertical slice", demo, or release; when estimating "how much before the demo"; when someone wants to "set up this game project" or implement an approved piece of game work; when running long or autonomous game-development sessions; or when taking Godot or browser-game work from an approved milestone to a checked playable build. NOT for discovering or defining direction (use game-direction), judging "is this done" or other readiness claims (use game-review), or engine-only debugging, upgrades, and maintenance detached from a playable build.

0
installs
▌
Credo Taste · gigio1023 bundle
Think through a long-horizon research decision one question at a time: which problem is worth months, whether to start or continue a project (best-case conclusion, one idea, riskiest sub-problem, continue / kill / pivot), what a new system or dataset could break, and how to read a rejection. Use for 'is this worth it', 'should I start this', 'this isn't working', 'is there an attack here'. Reasons in prose, no scores. NOT for this week's deliverables or tickets, for planning paper text (credo-paper-plan), or for reading a paper (credo-read-then-forget).

0
installs
▌
Credo Journal · gigio1023 bundle
Keep the record that trains research taste and run the monthly review on it: append ideas (problems worth solving), predictions made at decisions, hindsight when outcomes arrive, and a monthly review of the long-horizon portfolio (which one or two efforts could be great, what dies, months-ahead guesses). Use when an idea appears, a decision is made, a result or review comes back, or monthly and for 'what should I be working on'. Append-only through the bundled script. NOT for task tracking, meeting notes, or near-term deliverables.

0
installs
▌
Credo Release · gigio1023 bundle
Check a LaTeX paper before a conference submission, camera-ready, or arXiv upload in two separated phases: a bundled script for everything a program can catch (TODO markers, merge-conflict markers, doubled words, '..' and '??', author names under blind review, abstract macros, commented-out acknowledgments, build-log errors), then the human items the author confirms one by one (title, layout orphans, captions, fonts, page limits, anonymity beyond names, author block, copyright, template, arXiv form). Use for 'submit', 'camera-ready', 'arXiv', 'final check', 'lint the tex'. NOT for content or structure review (credo-paper-plan).

0
installs
▌
Credo Paper Plan · gigio1023
Plan a research paper before any prose is written: the one idea, the reader and what they believe now, the best-case conclusion, the introduction's story arc, a figure storyboard, and why each section sits where it does; then critique a draft against that plan. Use when the user is about to write a paper or section, or has a draft that reads as two papers. The agent plans and critiques; the author drafts. NOT for line editing, submission checks (credo-release), or short reports and tickets.

0
installs
▌
Credo Read Then Forget · gigio1023
Read a paper in a declared mode (scan for the one new thing, extract one needed piece, or reproduce), then name what the paper inherits from its field without argument and what the user would do if they had not read it. Use when a paper, preprint, or PDF lands and the user asks what it does, whether to read it, or what to take from it. NOT for surveys across many papers or related-work sections.

0
installs
▌
1password CLI · gigio1023 bundle
Use when the user wants to use the local macOS 1Password CLI (`op`) for setup or authentication, vault/item/document lookup or CRUD, secret references, `op read`, `op run`, `op inject`, service accounts, shell plugins, or SSH-key fields. Triggers on 1Password, `op`, vaults, items, OTP, CLI login, env-file secrets, 1Password shell plugins, or API tokens and SSH keys stored in 1Password. NOT for generic password advice, OpenClaw gateway/tmux auth flows, or silently replacing a CLI request with MCP, Connect, or Environments.

0
installs
▌
Skill Builder · gigio1023 bundle
Use when creating, improving, or auditing reusable agent skills: trigger descriptions, SKILL.md instructions, references, scripts, templates, packaging, or adaptation to a different model or harness. Turns real workflows and observed failures into scoped, verifiable packages. NOT for ordinary project documentation or one-off task prompts.

0
installs
▌
Codex Delegate · gigio1023 bundle
Use only from Claude Code, Cursor, or another non-Codex host harness when the user explicitly invokes or names codex-delegate to delegate a packaged, bounded task to the Codex CLI, or to resume, follow, or cancel such a run. Defaults to gpt-5.6-sol at xhigh effort on the standard non-Fast tier, with task-shaped routing to Terra or another model and effort when justified. Fast mode requires an explicit user request. Lets the host set Codex's judgment and internal-subagent authority in the packet. Uses a host-side launcher subagent as the default execution path when the harness supports one while the main host retains lifecycle ownership. Provides a durable detached run, file-only handoff, explicit sandbox, status, resume, cancellation, and a read-only renderer. NOT for automatic routing from mere mentions of Codex or GPT, use from inside Codex itself, official Codex plugin commands, general multi-agent orchestration, or delegating to Cursor (cursor-cli-delegation).

0
installs
▌
Goal Prompting · gigio1023 bundle
Use when the user explicitly wants a Codex Goal mode or Claude Code /goal prompt explained, drafted, reviewed, translated, set, or handed to the other harness. NOT for ordinary objectives, acceptance criteria, task prompts, implementation plans, or autonomous work without explicit Goal mode intent.

0
installs
▌
Frontend Design · gigio1023 bundle
Use for frontend work that can change what users perceive or operate: components, pages, layout, styling, responsive behavior, forms, data views, interface states, motion, accessibility, visual assets, and UI review. Includes user-visible frontend bug fixes. NOT for backend-only work or frontend configuration and data changes with no interface effect.

0
installs
▌
PDF Page Count · gigio1023 bundle
Use when the user asks for the page count of an existing PDF, whether a PDF meets an exact/minimum/maximum page limit, or the numeric page-count result for another workflow. Triggers on "PDF pages", "page count", "몇 페이지", and "분량 제한 확인". NOT for creating, editing, rendering, or visually reviewing a PDF; use the PDF skill for those tasks.

0
installs
▌
Mermaid Diagrams · gigio1023 bundle
Use when the user asks for Mermaid diagrams, flowcharts, sequence diagrams, architecture diagrams, diagram refactors, or diagram readability fixes. Guides type selection, reader-first layout, parser-safe syntax, palettes and dark-mode-safe styling, renderer compatibility, density control, and render validation. NOT for draw.io native XML authoring; use drawio-diagram.

0
installs
▌
Pr Review Comment · gigio1023 bundle
Use when the user wants code review findings validated against a Pull Request diff and posted as GitHub inline review comments, including "review and comment", "PR review comment", and "post review". Discovers an installed review skill or accepts supplied candidate findings, verifies code evidence and exact diff targets, previews the payload, and posts only after explicit approval. Works with github.com and GitHub Enterprise through the gh CLI. NOT for review-only feedback with no posting; use a code-review skill. NOT for validating or editing comments already posted on a PR.

0
installs
▌
Python Docstrings · gigio1023 bundle
Use when adding, correcting, auditing, or reviewing Python docstrings and explanatory comments, especially for public APIs, async and generator semantics, context-manager lifecycle, protocols and callbacks, side effects, exceptions, ordering, invariants, fallbacks, and compatibility constraints. Trigger for "docstring 추가", "주석 보강", "Python 주석", and "문서화해줘" when the target is Python source. NOT for general Markdown docs, annotation or API redesign, behavior-changing refactors, generated code, or comments that merely narrate statements.

0
installs
▌
Install Skill Pack · gigio1023 bundle
Use when installing or refreshing skills globally from a user-selected Git repository or revision with the Skills CLI for the standard Claude Code, Hermes Agent, OpenCode, Cursor, and Codex targets. Reviews packages and verifies revision identity, installed content, symlinks, and discovery. NOT for unpublished local changes, project-local installs, unscoped all-agent installs, or silently changing unrelated global skills.

0
installs
▌
Cross Harness Skills · gigio1023 bundle
Use when creating, reviewing, or modernizing one agent skill that must work in both Claude Code and Codex, especially when reconciling GPT-6 Astra, GPT-5.6-series, and Claude Fable prompting guidance, separating instructions from harness adapters, or testing the same skill across both runtimes. NOT for a single-harness extension or ordinary skill creation without a portability requirement; use skill-builder alone.

0
installs
▌
Toss Portfolio State · gigio1023 bundle
Use when the user wants a read-only Toss Invest OpenAPI portfolio snapshot, account balance/current holdings/order-history/market-context export, or normalized personal investment-state payload for another workflow. Triggers on "Toss API balance", "Toss OpenAPI holdings", "Toss portfolio snapshot", "Toss account-state export", "Toss market context", and "read my Toss investment status". NOT for investment advice, portfolio decisions, automatic trading, broker order creation/modification/cancelation, credential storage, or tax/legal conclusions; use investment-decision-support for judgment.

0
installs
▌
Fable5 Prompting Guide · gigio1023 bundle
Use when writing, reviewing, simplifying, debugging, or migrating system prompts, agent instructions, tool descriptions, long-run scaffolding, or prompt stacks specifically for Claude Fable 5. Applies current guidance for effort, strong instruction following, boundaries, progress evidence, subagents, memory, context, and refusals. NOT for general Anthropic API setup, model selection, or deciding whether Fable 5 should lead a task.

0
installs
▌
Cursor CLI Delegation · gigio1023 bundle
Use when the user explicitly invokes cursor-cli-delegation to delegate planning, research, implementation, review, commands, or verification through Cursor Agent CLI. Supports task-specific CLI options and related session follow-ups. NOT for automatic routing from Cursor/model mentions, or for launching Cursor merely to discuss or edit this skill.

0
installs
▌
English Prompt Review · gigio1023 bundle
Use when a user wants an English task prompt reviewed or quietly coached while the requested work continues, especially for software engineering, AI research, and game-development work. Rewrites the complete request in clear, natural technical English, explains nuance and ambiguity in precise Korean, and keeps key expressions and contrasts in English. NOT for translating non-English text, proofreading prose word by word, or changing the user's task requirements.

0
installs
▌
Gpt56 Sol Prompting Guide · gigio1023 bundle
Use when writing, reviewing, simplifying, debugging, or migrating developer prompts, system prompts, tool descriptions, agent instructions, or prompt stacks for GPT-5.6 Sol or the GPT-5.6 family. Applies current model-specific guidance to autonomy, tools, grounding, long runs, reasoning settings, and evals. NOT for choosing a model, general OpenAI API setup, or ordinary agent skill authoring without a GPT-5.6 prompt target.

0
installs
▌
Gpt6 Astra Prompting Guide · gigio1023 bundle
Use when writing, reviewing, debugging, or migrating system prompts, developer prompts, tool descriptions, agent instructions, skills, or prompt stacks specifically for GPT-6 Astra. Applies official OpenAI guidance to follow-through, clarification, instruction conflicts, writing style, delegation, proportional verification, and related runtime controls. NOT for model selection, general OpenAI API setup, GPT-5.6-only prompting, or ordinary skill authoring without an Astra prompt target.

0
installs
▌
Draft Pr · gigio1023 bundle
Use when the user wants to publish, update, or explicitly squash-merge an actual pull request on GitHub, Forgejo, or a compatible private forge, including "PR을 GitHub에 올려줘", "Forgejo에 draft PR을 열어줘", "open/create a PR", "submit this for review", "squash merge this PR", conflict cleanup, or an existing PR update. Requires explicit publication or merge intent; creates drafts or work-in-progress PRs by default, preserves reviewer context, and handles remote divergence safely. NOT for copy-only PR drafting, push-only requests handled by commit-and-push, issue workflows, or review comments.

0
installs
▌
Find Unknowns · gigio1023 bundle
Use when the user starts substantial work in territory they don't know well — a new project, research direction, investment or money-management strategy, career or life plan, product or game idea, or an unfamiliar part of a codebase — with no spec, plan, or reference yet; or when they ask for an unknowns pass, blind spot pass, option map, throwaway variants, or a launch brief. Surfaces unknowns with the single cheapest technique, compresses what was learned into a launch brief, and, when the user cannot test the result directly, explains consequential decisions and offers a comprehension check. NOT for well-specified or small tasks, multi-round Socratic discovery (use deep-interview), final decision synthesis (use fable5-model-routing), or packaging session state (use session-handoff).

0
installs
▌
Deep Interview · gigio1023 bundle
Use when the user explicitly wants to be interviewed deeply, grilled on a vague idea, led through requirements discovery, or questioned until a plan, product, workflow, or decision is clear enough to hand off. Runs an Ouroboros-centered, context-first Socratic interview one question at a time and ends with a user-approved Interview Brief. NOT for job-interview rehearsal, surveys, therapy or clinical intake, a single clarification, requests to summarize existing material without further questions, or light discovery that one technique plus a launch brief would settle (use find-unknowns).

0
installs
▌
Commit And Push · gigio1023
Use when the user explicitly asks to commit, split commits, stage and commit, push, sync a branch, or commit directly to the default branch. Triggers on "commit", "push", "git push", "커밋", "분할 커밋", "push 해줘", "stage and commit", and "direct commit". Keeps unrelated worktree changes untouched and limits the default result to commits plus branch push. NOT for opening or rewriting a PR (use draft-pr), full CI certification, or issue/PR comments and edits unless the user explicitly includes them.

0
installs
▌
Session Handoff · gigio1023 bundle
Use only when the user asks for a successor-ready handoff prompt, continuation prompt, next-agent prompt, or session transfer file, asks to package current work so another agent can continue without rereading the whole conversation, or names session-handoff. Creates one executable prompt file from intent, decisions, artifacts, live state, evidence, and remaining work. NOT for bounding a weaker executor on an already-approved plan (use small-model-handoff), or for writing or executing plans (gigio-write-plan, gigio-execute-plan). Never activate because a session has grown long or context is running low.

0
installs
▌
Use Terminology · gigio1023 bundle
Use on every task alongside curate-terminology, even when the user does not ask about wording. Read and apply accepted English terms, verbs, collocations, sentence patterns, company names, local meanings, and project explanations throughout responses and artifacts. NOT a standalone research workflow: curate-terminology owns source selection, new definitions, expression records, and corrections to the shared reference.

0
installs
▌
Gigio Write Plan · gigio1023 bundle
Use only when the user asks for a plan file or a revision of one, or names gigio-write-plan: produces one file in .plans/ with staged tasks — needs, owned files, acceptance, checks — anchored to PROJECT.md judgment rules. The plan is data: a solo session reads it top to bottom, a lead uses the same file to dispatch parallel workers. Writes the plan, announces it, and stops — it never implements tasks itself. NOT for executing or resuming a plan (gigio-execute-plan), reviewing finished work (gigio-review-results), or discovery while the work itself is still unclear (find-unknowns, deep-interview). Never activate because work looks sizable; ordinary in-session task tracking is not this skill and needs no plan file.

0
installs
▌
Curate Terminology · gigio1023 bundle
Use on every task alongside use-terminology, even when the user does not ask about wording. Maintain terminology and expressions as relevant material is encountered: define, research, update, and record accepted names, verbs, collocations, explanations, and anti-patterns from established projects and influential papers, plus user-confirmed company names and local meanings. NOT a requirement to run a fresh literature survey or write files when nothing relevant has changed; use-terminology applies the accepted reference to the current work.

0
installs
▌
Gigio Execute Plan · gigio1023
Use only when the user asks to execute or resume a plan file from .plans/ — "run the plan", "continue", "이어서 진행해줘", a plan path, a task ID, a stage name, or the skill name. Verifies the plan against the working tree first, dispatches independent same-stage tasks through available workers, judges completion from files and fresh check output rather than worker reports, and appends a run log. Resuming after an interruption is the same entry point invoked again. NOT for writing or revising plans (gigio-write-plan) or reviewing finished work against intent (gigio-review-results); when the plan itself turns out wrong, stop and route back to gigio-write-plan instead of patching it here. Never activate because a plan file exists — the user starts the run.

0
installs
▌
Git Worktree Setup · gigio1023 bundle
Use when implementation should be isolated from the current checkout, a plan calls for a separate branch workspace such as `gigio-execute-plan` isolating parallel workers, or worktree state must be detected before editing. Reuses existing isolation or creates a worktree through the harness or Git. NOT for merely listing worktrees, deleting them, or publishing a finished branch.

0
installs
▌
Write Internal Doc · gigio1023 bundle
Use only when the user asks for a document that colleagues will read without this session's context: a shared report, handbook, onboarding page, research write-up, decision or status memo, a Notion page, a shareable HTML file, or a shared copy of an existing document (사내 공유 문서, 리포트, 핸드북, 온보딩 문서, 조사 보고서, 결정 메모, 공유본, 노션에 올릴 문서), or names write-internal-doc. Shapes the document around one reader, keeps every claim traceable to a source a colleague can open, states what the records cannot show, runs a sharing pass before anything leaves the machine, and hands prose, Korean, figures, and Notion styling to the skills that own them. NOT for chat answers, plan files (gigio-write-plan), next-session prompts (session-handoff), PR or commit copy, external client deliverables, or prose polish alone (slop-aware-writing). Never activate because a conversation produced findings worth sharing or because an answer ran long.

0
installs
▌
Gigio Project Setup · gigio1023
Use only when the user asks to install the gigio-pack system into a project or audit an existing installation, or names gigio-project-setup: writing PROJECT.md (diagnosis, pillars, non-goals, judgment rules, current risk, current position), wiring shared agent instructions into AGENTS.md with a CLAUDE.md bridge, or checking an installed PROJECT.md for staleness. Triggers on adopting this pack in a repo, "set up the project intent file", "PROJECT.md 만들어줘/점검해줘". NOT for writing task plans (gigio-write-plan), executing them (gigio-execute-plan), reviewing finished work (gigio-review-results), or surfacing unknowns before the work itself is chosen (find-unknowns). Never activate because a repository has no PROJECT.md.

0
installs
▌
Small Model Handoff · gigio1023 bundle
Use only when the user invokes small-model-handoff by name, or when a pack skill (gigio-execute-plan, orchestrate-subagents) name-calls it while dispatching to a weaker executor, to turn an already researched and approved plan into a bounded prompt for an executor less capable than the planner. Covers scoped changes, test, build, and check runs, reproductions, and read-only evidence collection. Requires exact scope, actions, authority, capability assumptions, evidence, and stop rules. NOT for automatic invocation, model/cost/executor recommendations, peer or high-capability executors, discovery, planning, open-ended work, or general session handoffs.

0
installs
▌
Fable5 Model Routing · gigio1023 bundle
Claude Code and Cursor only. Use only when the user names fable5-model-routing or asks to put Fable 5 on the judgment — difficult judgment, long-horizon work, issue review, decision synthesis, overlooked-fact checks, strategic recommendations, or critique. Then inspect the actual problem and independently confirm that it has a consequential, unresolved judgment core for which Fable 5 has a concrete advantage; the request itself is not that confirmation. Do not use in Codex, when Fable 5 is unavailable, or for routine bounded work the current agent can finish. Never activate because the task at hand looks hard.

0
installs
▌
Gigio Review Results · gigio1023
Use only when the user asks for finished or long-running work to be reviewed against the project's intent in fresh context, or names gigio-review-results — before closing a plan, after a long autonomous run, or on request. Re-collects the facts itself (diffs, files, re-run checks) instead of trusting execution reports, audits the plan item by item, and returns three lists: missing, built-but-not-asked, misunderstood. NOT for writing plans (gigio-write-plan), executing or resuming them (gigio-execute-plan), or ordinary code review of a change that has no plan or PROJECT.md to compare against. Never activate because a run just finished.

0
installs
▌
Orchestrate Subagents · gigio1023 bundle
Use only when the user asks for parallel subagents, delegated agents, agent teams, or competing research tracks, or names orchestrate-subagents. Orchestrates coding, research, judgment, review, planning, and synthesis across Codex, Claude Code, Cursor, OpenCode, Antigravity, and similar harnesses. NOT for small or tightly sequential tasks. Never activate because a task happens to have independent workstreams — spawning an agent or two inline needs no skill, and the size of the fan-out is the user's call.

0
installs
▌
Python Coding Standards · gigio1023 bundle
Use when writing, refactoring, or reviewing Python application or library code, defining data models, splitting a large Python module, or setting up a Python project. Prefers straightforward Pydantic models, explicit Python 3.12+ annotations including locals, uv project management, and explanations close to code. Uses the official external pydantic skill for library guidance. NOT for docstring-only edits, a standalone formatting command, generated code unless explicitly in scope, or starting a plan or audit merely because Python files exist; gigio-review-results owns comparison of finished work against project intent.

0
installs

51 published skills