feature-planning-artifacts
Use when creating or iterating staged planning artifacts for complex features: design discussions, structure outlines, final execution plans, or exploratory…
Skill by TheCheepeer
Install
Works with Claude Code, Claude apps, Codex, Cursor, Gemini CLI, GitHub Copilot, VS Code, OpenCode, Goose, Windsurf. Read the repository before you give any tool access to your data.
npx agentmarket install skill/thecheepeer-feature-planning-artifacts
Asks which of your tools to add it to (Claude Code, Codex, Cursor, VS Code, Windsurf, Gemini CLI, Claude Desktop) and shows the change first. See https://agentmart.dev/cli.
Get the weekly diff
The weekly diff: one email a week with new agent tools, notable updates and grade changes, plus security alerts for tools you follow. Double opt-in, one-click unsubscribe.
About
Use cases: design, research
Reviews
Write a reviewNo reviews yet. Installed it? Reviews come from verified installers only. How reviews work
Permissions
What this skill can do and what it asks for, from its SKILL.md. Checked Oct 25, 2026.
- Tools
- No tools found
- Secrets
- None asked for
- Install scripts
- None found
Network hosts it talks to 1
codeload.github.com· found in install command
Files it touches
~/.claude/skills/feature-planning-artifacts
What's inside
- Skill name
feature-planning-artifacts- Description
- Use when creating or iterating staged planning artifacts for complex features: design discussions, structure outlines, final execution plans, or exploratory technical research. Slices work into executable vertical layers, establishes review gates with the user between stages, and derisks high-value items before code is written.
- License
- Apache-2.0
- Instructions
- 76 lines in SKILL.md · view on GitHub
Files (3)
- README.md
- SKILL.md
- artifact-templates.md · references/
SKILL.md (first lines)
# Feature Planning Artifacts Create or update structured, staged planning artifacts for requirements demanding prior research, architectural judgment, vertical slicing, and a verified plan before implementation. This skill is **not** a throwaway, single-pass plan generator. It takes a scope, performs reconnaissance, drafts or updates the current-stage artifact, and **stops at the review gate**. Do not advance to the next stage until the current artifact is approved by the user or they explicitly request proceeding. Treat planning as meta-engineering: each document makes future implementation safer by specifying success criteria, clarifying architectural decisions, and designing regression gates. ## Sources and References Before producing or updating artifacts, consult: - The selected roadmap item, opportunity card, or preliminary brief. - Prior artifacts for this effort (design discussion, outline, or plan drafts). - The `coding-standards` skill as the canonical source for vocabulary and code idioms. - Project documents: `GEMINI.md`, `AGENTS.md`, `README.md`, ADRs, and current Git status. - Canonical templates in [references/artifact-templates.md](references/artifact-templates.md). ## The Stage Contract Always advance via the smallest stage that yields solid progress: | Stage | When to Use | Input | Resulting Output | Stop Gate | | ------------------------ | --------------------------------------------------------- | ----------------------------------------- | ---------------------------------------------------------- | -------------------------------------------------- | | **Research Questions** | Broad uncertainty about how the system currently works | User request or roadmap item | Precise questions on current code behavior | Stop after questions; await approval to research | | **Technical Research** | Raised questions must be answered by the code | Research questions | Purely descriptive report of current code | Stop after research; recommend design discussion | | **Design Discussion** | The feature requires architectural choices and trade-offs | Research or sufficient repository context | Options, recommendation, settled decisions, open questions | Stop for human review; DO NOT draft outline yet | | **Structure Outline** | Design is approved or has a clear direction | Accepted design discussion | Vertical implementation slices | Stop for human review; DO NOT draft final plan yet | | **Final Execution Plan** | Structure outline is approved | Accepted outline | Self-contained, safe execution plan for any executor | Stop for execution handoff | If the user simply says "plan this", begin with a **design discussion**, unless uncertainties are so extensive that research questions must come first. Never generate everything at once without pauses. ## File Structure and Destination Colocate the artifact bundle in `.agents/plans/features/<feature-slug>/`: ```txt .agents/plans/features/<feature-slug>/README.md .agents/plans/features/<feature-slug>/001-design-discussion.md .agents/plans/features/<feature-slug>/002-structure-outline.md .agents/plans/features/<feature-slug>/003-plan.md ``` If descriptive research was required: ```txt .agents/plans/features/<feature-slug>/001-research-questions.md .agents/plans/features/<feature-slug>/002-research.md .agents/plans/features/<feature-slug>/003-design-discussion.md ``` Create or update exactly **one primary artifact** per interaction, along with the bundle index `README.md`. ## Detailed Workflow by Stage 1. **Define the Current Stage:** Assess existing files. If a design discussion is under review, iterate on it rather than creating an outline. Define success criteria before writing. 2. **Technical Research:** Keep research strictly descriptive (current code behavior, data flows, tests, and constraints). No solution recommendations or opinions in this phase. 3. **Design Discussion:** Present current state, desired end state, explicit non-goals, evaluated alternatives, technical recommendation, and open questions. Set `status: in-review` and request user validation. 4. **Structure Outline:** Once design is accepted, slice implementation into vertical phases. List touched files, core signatures, and per-phase tests. 5. **Final Plan:** Derived from the accepted outline, draft an executor-ready plan with exact test commands, stop conditions, and autonomy boundaries. ## User Review Gates Use clear, direct handoff prompts when concluding a stage: - **After research questions:** _"These are the questions regarding current system behavior I will investigate next."_ - **After research:** _"Here is the descriptive mapping of existing code; solution recommendations will be detailed in the design discussion."_ - **After design discussion:** _"Please review the recommended approach, discarded options, and open questions before we outline implementation phases."_ - **After structure outline:** _"Please review the vertical delivery slices and phase-level tests before we generate the detailed execution plan."_ - **After final plan:** _"The plan is complete and ready for safe execution."_ ## Completion Criteria The planning bundle is complete when each document has fulfilled its role and cleared its validation gate, allowing the feature to be implemented without hidden questions or architectural friction.
From TheCheepeer/omni-agents, Apache-2.0.
More from TheCheepeer/omni-agents
- 0.570score
strategic-roadmap
TheCheepeer
Use when the user asks "what should I work on in this repository?", "what is worth doing next?", requests a repository audit, technical debt priorities…
SkillAFreshness: AClaude Code +9 - 0.570score
writing-error-messages
TheCheepeer
Use when writing, reviewing, or rewriting human-facing error messages, form validation errors, empty states, authentication failures, error notifications…
SkillAFreshness: AClaude Code +9 - 0.570score
frontend-design-principles
TheCheepeer
Use when building or reviewing frontend user interfaces (UI) — dashboards, admin panels, landing pages, marketing sites, and web applications.
SkillAFreshness: AClaude Code +9 - 0.570score
html-artifacts
TheCheepeer
Use when asked to create, generate, visualize, or convert material into a standalone HTML artifact, single-file self-contained .html report, interactive…
SkillAFreshness: AClaude Code +9 - 0.570score
coding-standards
TheCheepeer
Use when reviewing code quality, evaluating whether a refactoring or design is sound, identifying code smells, or judging architectural trade-offs.
SkillAFreshness: AClaude Code +9 - 0.570score
reducing-entropy
TheCheepeer
Use when evaluating architectures, reviewing code, or planning refactorings — measures success by total code volume in the end state rather than immediate…
SkillAFreshness: AClaude Code +9 - 0.570score
researching-codebases
TheCheepeer
Use when answering complex questions about a codebase that require exploring multiple modules or tracing flow across components.
SkillAFreshness: AClaude Code +9 - 0.570score
improving-prompts
TheCheepeer
Use when optimizing GEMINI.md, AGENTS.md, system prompts, custom slash commands, or skill files.
SkillAFreshness: AClaude Code +9 - 0.570score
skill-authoring
TheCheepeer
Use when creating, authoring, refining, or debugging skills for AI agents.
SkillAFreshness: AClaude Code +9 - 0.570score
writing-cli-skills
TheCheepeer
Use when authoring an agent skill that wraps a command-line interface (CLI) tool or terminal binary.
SkillAFreshness: AClaude Code +9 - 0.570score
crafting-effective-readmes
TheCheepeer
Use when drafting, structuring, or improving README.md files.
SkillAFreshness: AClaude Code +9 - 0.570score
diataxis
TheCheepeer
Structure, classify, and author technical documentation using the Diátaxis framework.
SkillAFreshness: AClaude Code +9
Signals
- Stars
- 0
- Forks
- 0
- Open issues
- 0
- Last commit (this folder)
- Sep 28, 2026 (1 month ago)
From GitHub (fetched Oct 24, 2026).
Details
- License
- Apache-2.0
- Repository
- github.com/TheCheepeer/omni-agents
- First seen
- Oct 24, 2026
- Last checked
- Oct 24, 2026
- Source
- GitHubgithub-graphql
More from TheCheepeer
- 0.570score
writing-error-messages
TheCheepeer
Use when writing, reviewing, or rewriting human-facing error messages, form validation errors, empty states, authentication failures, error notifications…
SkillAFreshness: AClaude Code +9 - 0.570score
frontend-design-principles
TheCheepeer
Use when building or reviewing frontend user interfaces (UI) — dashboards, admin panels, landing pages, marketing sites, and web applications.
SkillAFreshness: AClaude Code +9 - 0.570score
html-artifacts
TheCheepeer
Use when asked to create, generate, visualize, or convert material into a standalone HTML artifact, single-file self-contained .html report, interactive…
SkillAFreshness: AClaude Code +9 - 0.570score
coding-standards
TheCheepeer
Use when reviewing code quality, evaluating whether a refactoring or design is sound, identifying code smells, or judging architectural trade-offs.
SkillAFreshness: AClaude Code +9 - 0.570score
reducing-entropy
TheCheepeer
Use when evaluating architectures, reviewing code, or planning refactorings — measures success by total code volume in the end state rather than immediate…
SkillAFreshness: AClaude Code +9 - 0.570score
writing
TheCheepeer
Prose and writing guidelines for human-facing technical communication — clear, concrete, and free from prefabricated AI clichés.
SkillAFreshness: AClaude Code +9
Cite this page
feature-planning-artifacts skill. Agent Mart. https://agentmart.dev/skills/thecheepeer-feature-planning-artifacts (last updated 2026-10-24)