design-process
7-phase agent design process — DISCOVER, SCOPE, SPECIFY, REVIEW & APPROVE, GENERATE, VALIDATE, PERSIST.
Skill by edgeai
Chalk Labs
Need help deploying design-process?
Chalk Labs sets up agents like this inside real teams: connected to your tools, tested on your data, with a human in the loop where it matters.
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/pedrogrande-design-process
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, agent-building
Reviews
Write a reviewNo reviews yet. Installed it? Reviews come from verified installers only. How reviews work
Permissions: not checked yet.
What's inside
- Skill name
design-process- Description
- 7-phase agent design process — DISCOVER, SCOPE, SPECIFY, REVIEW & APPROVE, GENERATE, VALIDATE, PERSIST. Load when designing a new agent or continuing an in-progress design.
- License
- MIT
- Metadata
- author: edgeai · version: 1.0.0
- Instructions
- 37 lines in SKILL.md · view on GitHub
Files (8)
- SKILL.md
- discover.md · references/
- generate.md · references/
- persist.md · references/
- review-approve.md · references/
- scope.md · references/
- specify.md · references/
- validate.md · references/
SKILL.md (first lines)
# Agent Design Process
You follow a 7-phase process for every agent design. Phases 1–3 are conversational.
Phase 4 is a mandatory human review gate. Phases 5–7 only proceed after approval.
## Phase Overview
| Phase | Name | Key Activity | Gate? |
|-------|------|-------------|-------|
| 1 | DISCOVER | Understand purpose, users, cognitive mode | No |
| 2 | SCOPE | Architecture, tools, knowledge, memory, storage | No |
| 3 | SPECIFY | Model, prompt, tools, knowledge, memory, guardrails | No |
| 4 | REVIEW & APPROVE | Present completed template for human approval | **Yes** |
| 5 | GENERATE | Produce complete Python agent file + setup instructions | No |
| 6 | VALIDATE | Verify imports, params, env vars, dependencies | No |
| 7 | PERSIST | Save spec to database, advance lifecycle status | No |
## When to Load Sub-Skills
Load the reference for the phase you're currently in:
- **Phase 1** → `get_skill_reference("design-process", "discover.md")`
- **Phase 2** → `get_skill_reference("design-process", "scope.md")`
- **Phase 3** → `get_skill_reference("design-process", "specify.md")`
- **Phase 4** → `get_skill_reference("design-process", "review-approve.md")`
- **Phase 5** → `get_skill_reference("design-process", "generate.md")`
- **Phase 6** → `get_skill_reference("design-process", "validate.md")`
- **Phase 7** → `get_skill_reference("design-process", "persist.md")`
## Cross-Phase Rules
1. **Never skip Phase 4** — the completed template MUST be approved before code generation.
2. **Check session state before requesting approval** — if the spec was already approved and nothing changed, don't ask again.
3. **Use your knowledge base FIRST** for API verification — only fall back to DuckDuckGo when the KB has no results.
4. **Explain technical choices in plain language** — load the `non-technical-explanations` skill when the user is non-technical.
5. **Always use Ollama models** — default `Ollama(id="glm-5.1:cloud")`. Embeddings are the only exception (use `OpenAIEmbedder`).From pedrogrande/adf, MIT.
Signals
- Stars
- 0
- Forks
- 0
- Open issues
- 0
- Last commit (this folder)
- May 11, 2026 (5 months ago)
From GitHub (fetched Oct 24, 2026).
Details
- Version
- 1.0.0
- License
- MIT
- Repository
- github.com/pedrogrande/adf
- First seen
- Oct 24, 2026
- Last checked
- Oct 24, 2026
- Source
- GitHubgithub-graphql
Same skill elsewhere
Copies of this SKILL.md published by others; search shows only one of them.
More from edgeai
- 0.486score
design-template
edgeai
Full 8-part Agno Agent Design Template.
SkillCFreshness: C★ 1Claude Code +9 - 0.429score
architecture-context
edgeai
Current EdgeAI infrastructure reference — databases, agents, tools, services, and environment variables.
SkillDFreshness: D★ 1Claude Code +9 - 0.429score
design-process
edgeai
7-phase agent design process — DISCOVER, SCOPE, SPECIFY, REVIEW & APPROVE, GENERATE, VALIDATE, PERSIST.
SkillDFreshness: D★ 1Claude Code +9 - 0.424score
non-technical-explanations
edgeai
Plain-language analogies for technical design choices.
SkillDFreshness: D★ 1Claude Code +9
Done for you
Rather not set it up yourself?
Chalk Labs scopes, builds and runs agents for your team: the task audit, the eval harness, the integrations and the weekly ops report. Fixed scope, priced before we start.
See how we build AI agents →Read the case studies
Chalk Labs is a two-founder AI and growth studio. We build AI agents, CRMs and MVPs for Web3 and AI companies, and have shipped 50+ product and token launches.
- launches shipped
- 50+
- launches shipped
- media placements
- 500+
- media placements
- founders on every account
- 2
- founders on every account
Press we've placed: Forbes · CoinDesk · Decrypt · Cointelegraph · The Block
Rahil Jain, Co-founder, Chalk Labs. Ex-Goldman Sachs engineer. You talk to the person who builds it.
Get design-process running in your business
Tell us what you want to automate. A founder replies within one working day with a plan, not a pitch deck.
Cite this page
design-process skill. Agent Mart. https://agentmart.dev/skills/pedrogrande-design-process (last updated 2026-10-24)