Skip to content

Commit 8e7c08b

Browse files
jamesmcroftjamesmcroft
committed
Add spec-driven development generator for coding agents
1 parent 4d79ace commit 8e7c08b

File tree

3 files changed

+233
-9
lines changed

3 files changed

+233
-9
lines changed

‎README.md‎

Lines changed: 38 additions & 9 deletions
Original file line numberDiff line numberDiff line change
@@ -10,15 +10,44 @@ A library of copy-ready prompts for ChatGPT, Copilot, and similar tools. Every p
1010

1111
All prompts live under [`prompts/`](prompts/README.md). The folder categories are intentionally opinionated to keep things organized:
1212

13-
| Domain | Folder | Highlights |
14-
| ---------------------- | --------------------------------- | ------------------------------------------------------------- |
15-
| Content - Writing | `prompts/content/` | Outlining, SEO metadata, feature images, conclusions |
16-
| Content - Reviewing | `prompts/content/reviewing/` | Language quality checks, topic accuracy, readability feedback |
17-
| Social Media | `prompts/social/` | Conversation-starting posts and image prompts |
18-
| Coding | `prompts/coding/` | Release notes plus GitHub bug-report templates |
19-
| Knowledge Retrieval | `prompts/knowledge/` | Context-grounded answering for RAG setups |
20-
| Business Communication | `prompts/business/communication/` | Email drafting, stakeholder messaging helpers |
21-
| Career Development | `prompts/career/development/` | Performance reviews, growth plans, and professional coaching |
13+
### Coding
14+
15+
| Category | Highlights |
16+
| ------------------------------------------------------------------- | -------------------------------------------- |
17+
| [`Release Management`](prompts/coding/release-management/README.md) | Release notes and changelog prep |
18+
| [`Issue Management`](prompts/coding/issue-management/README.md) | GitHub bug reports and issue triage |
19+
| [`Coding Agents`](prompts/coding/coding-agents/README.md) | Spec-driven work items for autonomous agents |
20+
21+
### Business
22+
23+
| Category | Highlights |
24+
| ----------------------------------------------------------- | --------------------------------------------- |
25+
| [`Communication`](prompts/business/communication/README.md) | Email drafting, stakeholder messaging helpers |
26+
27+
### Career
28+
29+
| Category | Highlights |
30+
| ----------------------------------------------------- | -------------------------------------------------------- |
31+
| [`Development`](prompts/career/development/README.md) | Performance reviews, growth plans, professional coaching |
32+
33+
### Content Writing
34+
35+
| Category | Highlights |
36+
| -------------------------------------------------------------- | ------------------------------------------------------------- |
37+
| [`Article Writing`](prompts/content/article-writing/README.md) | Outlining, SEO metadata, feature images, conclusions |
38+
| [`Reviewing`](prompts/content/reviewing/README.md) | Language quality checks, topic accuracy, readability feedback |
39+
40+
### Social Media
41+
42+
| Category | Highlights |
43+
| ----------------------------------------------------------- | --------------------------------------------- |
44+
| [`Media Creation`](prompts/social/media-creation/README.md) | Conversation-starting posts and image prompts |
45+
46+
### Knowledge
47+
48+
| Category | Highlights |
49+
| ------------------------------------------------------------- | ----------------------------------------- |
50+
| [`Retrieval & Q&A`](prompts/knowledge/retrieval-qa/README.md) | Context-grounded answering for RAG setups |
2251

2352
## How to Use
2453

‎prompts/coding/coding-agents/README.md‎

Lines changed: 8 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,8 @@
1+
---
2+
title: Coding Agents
3+
description: Prompts that support agentic workflows, from planning specs to orchestrating autonomous coding tasks.
4+
---
5+
6+
# Coding Agents
7+
8+
- [Spec-Driven Development Generator](spec-driven-work-item.md) - produce full feature specs (context, requirements, Gherkin criteria, effort, constraints) ready for coding agents to execute.

‎prompts/coding/coding-agents/spec-driven-work-item.md‎

Lines changed: 187 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,187 @@
1+
---
2+
title: Spec-Driven Development Generator
3+
category: coding
4+
subcategory: coding-agents
5+
description: Transform a feature idea into a full technical spec that coding agents can execute autonomously.
6+
llm_tools:
7+
- GitHub Copilot
8+
- ChatGPT Codex
9+
- Claude Code
10+
inputs:
11+
- Feature or change description
12+
- Any known context or files to inspect
13+
---
14+
15+
# Spec-Driven Development Generator
16+
17+
Use this prompt inside GitHub Copilot, ChatGPT Codex, or Claude Code ask mode to produce a spec that can be either automatically acted on in said agent, or copied into an Azure DevOps, GitHub Issues, or other trackers before handing the task to an autonomous coding agent.
18+
19+
## Prompt
20+
21+
```markdown
22+
You are a spec-driven development work item generator. Your task is to take a user's description of a desired feature, change, or enhancement, and convert it into a robust technical specification suitable for guiding development efforts.
23+
24+
When the user provides the description, you must analyze the codebase for context and output a complete, structured specification containing the following details:
25+
26+
---
27+
28+
## **1. Context / Problem Statement**
29+
30+
Describe the background, current behavior, and why this change is needed.
31+
Include references to existing systems, workflows, dependencies, technical context, and architectural considerations.
32+
33+
---
34+
35+
## **2. High-Level Overview**
36+
37+
A clear summary of what is being added, changed, removed, or enhanced.
38+
39+
---
40+
41+
## **3. Detailed Requirements**
42+
43+
Provide a numbered list of explicit, testable requirements.
44+
Include both:
45+
46+
- **Functional Requirements (FR):** Intended behaviors, system actions, expected inputs/outputs, workflows, data handling rules.
47+
- **Non-Functional Requirements (NFR):** Performance, reliability, security, UX, availability, compliance, AI agent behavior, guardrails, latency, cost considerations, etc.
48+
49+
---
50+
51+
## **4. Expected Functionality**
52+
53+
Explain how the system/agent/application should behave after implementation.
54+
Include:
55+
56+
- Step-by-step operation flow
57+
- User interaction flow (if relevant)
58+
- Agent orchestration behavior (if relevant)
59+
- Integration behavior with APIs, services, models, environments
60+
- Edge cases and fallback behavior
61+
62+
---
63+
64+
## **5. Acceptance Criteria (Gherkin Style)**
65+
66+
Write acceptance criteria using **Given / When / Then** format.
67+
Include:
68+
69+
- Core success cases
70+
- Error conditions
71+
- Edge cases
72+
- Agent-specific behaviors
73+
- Automation test criteria (if applicable)
74+
75+
---
76+
77+
## **6. Priority Level (0–4)**
78+
79+
- **0 – Critical** (blocks users or system, must be done immediately)
80+
- **1 – High** (major value; required for core workflows)
81+
- **2 – Medium** (important but not urgent)
82+
- **3 – Low** (nice-to-have)
83+
- **4 – Icebox** (future idea / backlog)
84+
85+
---
86+
87+
## **7. Effort Estimate**
88+
89+
Provide a relative effort estimate measured in hours. Include justification based on expected complexity.
90+
91+
---
92+
93+
## **8. Constraints**
94+
95+
List all constraints, such as:
96+
97+
- Technical limitations
98+
- API rate limits
99+
- Security/compliance requirements
100+
- Model behavior constraints
101+
- Environmental constraints (e.g., Azure Config, container limits)
102+
- Legacy system boundaries
103+
- Performance expectations
104+
- Dependency interactions
105+
106+
---
107+
108+
## **9. Dependencies & Integration Points**
109+
110+
Include:
111+
112+
- Upstream/downstream systems
113+
- Data sources
114+
- External services
115+
- Microservices / agents / pipelines
116+
- Versioning constraints
117+
- Required domain knowledge
118+
119+
---
120+
121+
## **10. Open Questions / Unknowns**
122+
123+
A short list of clarifications that need resolution before coding begins.
124+
125+
---
126+
127+
## **11. Output Format**
128+
129+
All outputs must be delivered as Markdown using this structure:
130+
131+
<STRUCTURE>
132+
# Feature / User Story Specification
133+
134+
## 1. Context / Problem Statement
135+
136+
...
137+
138+
## 2. High-Level Description
139+
140+
...
141+
142+
## 3. Detailed Requirements
143+
144+
### 3.1 Functional Requirements
145+
146+
1.
147+
148+
### 3.2 Non-Functional Requirements
149+
150+
1.
151+
152+
## 4. Expected Functionality
153+
154+
...
155+
156+
## 5. Acceptance Criteria
157+
158+
- **Scenario:**
159+
**Given**
160+
**When**
161+
**Then**
162+
163+
## 6. Priority
164+
165+
...
166+
167+
## 7. Effort
168+
169+
...
170+
171+
## 8. Constraints
172+
173+
...
174+
175+
## 9. Dependencies
176+
177+
...
178+
179+
## 10. Open Questions
180+
181+
...
182+
183+
</STRUCTURE>
184+
```
185+
186+
> [!TIP]
187+
> Once the spec is generated in chat/ask mode, switch to Agent Mode so Copilot or another coding agent can implement it using the spec as the source of truth, or paste it into a work item.

0 commit comments

Comments
 (0)