Exécutez n'importe quel Skill dans Manus
en un clic

Exécutez n'importe quel Skill dans Manus en un clic

grill-with-docs

Étoiles3

Forks0

Mis à jour10 juin 2026 à 07:28

Stress-test De-Koi plans against the current repo, product docs, architecture rules, and durable guidance, then capture only reusable decisions. Use when the user wants to be grilled on a design, sharpen fuzzy requirements, validate terminology, or update docs while a plan is clarified.

Installation

Installer avec Codex ou Claude Copiez ce prompt, collez-le dans Codex, Claude ou un autre assistant, puis laissez-le vérifier la page du skill et l'installer pour vous.

Exécuter dans Manus

Source

The-Koi-Pond

The-Koi-Pond/De-Koi

Ouvrir le dépôt GitHub Voir les dépôts du créateur

Téléchargement

Exécuter dans Manus

Métiers associésSOC

Basé sur la classification professionnelle SOC

Spécialistes en gestion de projetsProfessions des affaires et des opérations financières·SOC 13-1082

Explorateur de fichiers

2 fichiers

SKILL.md

readonly

name	grill-with-docs
description	Stress-test De-Koi plans against the current repo, product docs, architecture rules, and durable guidance, then capture only reusable decisions. Use when the user wants to be grilled on a design, sharpen fuzzy requirements, validate terminology, or update docs while a plan is clarified.

Grill With Docs

Use this skill to reach shared understanding before implementation. Ask one question at a time. If code or docs can answer the question, inspect them instead of asking.

Load First

AGENTS.md
PRODUCT.md and DESIGN.md when product or UI intent matters
skills/de-koi-agent-workflow/SKILL.md
skills/de-koi-architecture-guard/SKILL.md for ownership, imports, shared APIs, Tauri, Rust, storage, providers, or runtime behavior
skills/de-koi-mode-separation/SKILL.md for chat, roleplay, game, prompt, generation, or mode UI behavior

Session Loop

Restate the plan in one concise paragraph.
Identify the highest-risk unclear decision.
Ask one question with your recommended answer.
Use repo inspection to resolve answerable questions.
Challenge fuzzy language against De-Koi terms and owners.
Probe concrete scenarios and edge cases.
Record only durable decisions that future work needs.
Continue until the plan has clear scope, owner, proof, and out-of-scope boundaries.

What To Challenge

owner confusion between src/engine, src/features, src/shared/api, and src-tauri
chat, roleplay, and game mode mixing
UI-only fixes for engine/storage/provider contract bugs
hidden persistence, migration, import/export, prompt, provider, or security risk
vague words such as "sync", "memory", "agent", "runtime", "profile", or "state" when multiple De-Koi meanings exist
missing negative controls for destructive or detection logic

Durable Capture

Do not create docs by reflex. Capture decisions only when they change future agent behavior, architecture, product language, or issue/PR clarity.

Use the narrowest durable home:

GitHub issue or PR body for active work ownership and acceptance criteria
skills/*/references/* for reusable agent guidance
developer docs for user-facing or architecture guidance that belongs outside skills

Do not add AI/tool self-attribution to public text. Draft exact external text and wait for approval unless posting was already authorized.

Plus depuis ce dépôt

même dépôt

de-koi-agent-workflow

The-Koi-Pond/De-Koi

Apply De-Koi's repo-local version of the Chai agent workflow pack for proof discipline, investigations, feature sizing, refactor handoffs, reviews, PR readiness, reviewer feedback iteration, issue drafting, UI/UX proof, security-sensitive checks, risky-work evidence, debt notes, and final status reports. Use for nontrivial coding work, ambiguous symptoms, refactors, reviews, PR/issue workflows, or any task where an agent must prove claims while preserving De-Koi's code separation and module ownership rules.

2026-06-123

de-koi-architecture-guard

The-Koi-Pond/De-Koi

Protect De-Koi's layered Tauri and hostable Rust runtime architecture, module ownership, dependency direction, explicit imports, Rust capability boundaries, shared-code placement, HTTP pipeline, and file-splitting discipline. Use when changing folders, imports, shared modules, TypeScript engine layers, Tauri/HTTP command wrappers, Rust capability crates, repositories, adapters, feature APIs, or any code structure that could widen impact area.

2026-06-113

de-koi-bugfix-discipline

The-Koi-Pond/De-Koi

Guide De-Koi bug fixes through feedback-loop-first diagnosis and root-cause repair with clear impact areas, focused validation, and no band-aid patches. Use for regressions, broken UI actions, failing checks, provider/transport bugs, storage bugs, import/export bugs, generation bugs, mode behavior bugs, performance regressions, or any change where the fix could affect dependent modules.

2026-06-103

improve-codebase-architecture

The-Koi-Pond/De-Koi

Find De-Koi architecture deepening opportunities that improve locality, leverage, testability, and agent navigation without violating repo boundaries. Use when the user asks to improve architecture, reduce shallow modules, plan refactors, make code more testable, or inspect module/interface design.

2026-06-103

prototype

The-Koi-Pond/De-Koi

Build throwaway De-Koi prototypes to answer logic, state-model, API-shape, or UI-layout questions before production implementation. Use when the user asks to prototype, mock up, sanity-check a state model, try UI variants, or let them play with a design.

2026-06-103

tdd

The-Koi-Pond/De-Koi

Guide De-Koi changes through behavior-first red-green-refactor cycles using public interfaces and repo proof rules. Use when the user asks for TDD, red-green-refactor, test-first development, regression tests, integration tests, or a risky behavior change that needs a committed proof guard.

2026-06-103

name	grill-with-docs
description	Stress-test De-Koi plans against the current repo, product docs, architecture rules, and durable guidance, then capture only reusable decisions. Use when the user wants to be grilled on a design, sharpen fuzzy requirements, validate terminology, or update docs while a plan is clarified.

Grill With Docs

Use this skill to reach shared understanding before implementation. Ask one question at a time. If code or docs can answer the question, inspect them instead of asking.

Load First

AGENTS.md
PRODUCT.md and DESIGN.md when product or UI intent matters
skills/de-koi-agent-workflow/SKILL.md
skills/de-koi-architecture-guard/SKILL.md for ownership, imports, shared APIs, Tauri, Rust, storage, providers, or runtime behavior
skills/de-koi-mode-separation/SKILL.md for chat, roleplay, game, prompt, generation, or mode UI behavior

Session Loop

Restate the plan in one concise paragraph.
Identify the highest-risk unclear decision.
Ask one question with your recommended answer.
Use repo inspection to resolve answerable questions.
Challenge fuzzy language against De-Koi terms and owners.
Probe concrete scenarios and edge cases.
Record only durable decisions that future work needs.
Continue until the plan has clear scope, owner, proof, and out-of-scope boundaries.

What To Challenge

owner confusion between src/engine, src/features, src/shared/api, and src-tauri
chat, roleplay, and game mode mixing
UI-only fixes for engine/storage/provider contract bugs
hidden persistence, migration, import/export, prompt, provider, or security risk
vague words such as "sync", "memory", "agent", "runtime", "profile", or "state" when multiple De-Koi meanings exist
missing negative controls for destructive or detection logic

Durable Capture

Do not create docs by reflex. Capture decisions only when they change future agent behavior, architecture, product language, or issue/PR clarity.

Use the narrowest durable home:

GitHub issue or PR body for active work ownership and acceptance criteria
skills/*/references/* for reusable agent guidance
developer docs for user-facing or architecture guidance that belongs outside skills

Do not add AI/tool self-attribution to public text. Draft exact external text and wait for approval unless posting was already authorized.