Install
$ agentstack add skill-realdougeubanks-claudemarketplace-requirements-generator ✓ scanned · ✓ verified, works with Claude Code, Cursor, and more.
Security review
✓ PassedNo issues found. Passed automated security review. · v0.1.0 How review works →
- ✓ Prompt-injection patterns
- ✓ Secret / credential exfiltration
- ✓ Dangerous shell & filesystem operations
- ✓ Untrusted network calls
- ✓ Known-malicious package signatures
What it can access
- ✓ Network access No
- ✓ Filesystem access No
- ✓ Shell / process execution No
- ✓ Environment & secrets No
- ✓ Dynamic code execution No
From automated source analysis of v0.1.0. “Used” means the capability is present in the source — more access means more to trust, not that it’s unsafe.
Verified badge
Passed review? Show it. Paste this badge into your README, it links to the public security report.
Reliability & compatibility
Declared compatibility
Compatibility is declared by the source manifest. End-to-end runtime verification is coming, see below.
We're building live execution health for every listing: tool-call success rate, median latency, uptime, and last-checked timestamps, measured, not self-reported. It isn't live yet, so we don't show numbers we can't stand behind.
How agent discovery & health will work →About
Requirements Generator
Take a project idea, feature description, or user story and generate a structured requirements document with functional requirements, non-functional requirements, Gherkin acceptance criteria, edge cases, and out-of-scope items.
Instructions
Two modes:
- Description mode (default): user provides a free-form description or pastes a Jira/Confluence URL. Claude generates requirements from that.
- Interview mode (
/requirements-generator --interview): Claude asks a structured set of questions one at a time, then generates requirements from the answers.
- Gather input. Ask the user what they want to build if no description has been provided. Accept any of:
- A pasted feature description or user story
- A Jira ticket URL or Confluence page URL (read via available MCP tools if present)
- A brief verbal description of the feature
Prompt: "Describe the feature or system you want to build. Include the user need it addresses and any known constraints."
If the user invoked with --interview, skip this prompt and proceed to the Interview Mode section below instead.
- Understand the existing codebase (if one exists). Use Glob on
*to detect the project structure. Readpackage.json,pyproject.toml, orgo.modto identify the tech stack. Read 3–5 representative files fromsrc/orapp/— entry points first (index.*,main.*,app.*), then one route/handler file and one model/schema file — to understand patterns and conventions. Do not read the whole source tree. This ensures requirements fit what already exists.
- Check for existing requirements. Use Glob on
docs/requirements/**/*.md. If matching files exist, read them to avoid duplicating requirements already captured. Note any related requirements in the new document.
- Determine the feature name. Derive a kebab-case feature name from the description (e.g., "user authentication" →
user-authentication). This becomes the output filename.
- Generate the requirements document with the following sections:
### Overview One paragraph summarizing what is being built, the user need it addresses, and why it matters.
### Stakeholders
A table of who uses this feature and what they need from it.
| Stakeholder | Role | Primary Need | |-------------|------|-------------| | End User | ... | ... |
### Functional Requirements Numbered list using the prefix FR-NNN (e.g., FR-001, FR-002). Each requirement must be:
- Specific and testable
- Implementation-agnostic (describe what, not how)
- Written in the form: "The system shall..."
### Non-Functional Requirements Cover all four dimensions:
- Performance: specific response time targets (e.g., "API responses under 200ms at p95")
- Security: authentication requirements, data sensitivity classification, encryption requirements
- Accessibility: WCAG AA compliance, keyboard navigation, screen reader support
- Scalability: expected concurrent users, request volume, data growth projections
### Acceptance Criteria Gherkin format (Given/When/Then) — one scenario per key functional requirement. Use this format exactly:
``gherkin Scenario: Given And When Then And ``
### Edge Cases & Error Scenarios Bulleted list covering:
- Invalid or malformed input
- System failures (database down, third-party API timeout)
- Concurrent access and race conditions
- Boundary conditions (empty state, maximum limits)
- Unauthorized access attempts
### Out of Scope Explicitly list what this feature does NOT include. This prevents scope creep during implementation. Be specific.
### Assumptions List any assumptions made during requirements generation. For each:
- State the assumption clearly
- Note what would change if the assumption is wrong
### Open Questions List questions that require stakeholder input before implementation can begin. For each question, note:
- Who is the right person to answer it
- What the impact is if it is not answered before development starts
- Write the document. Ensure
docs/requirements/exists (it may need to be created). Use Write to save the document todocs/requirements/-requirements.md.
- Offer Jira/Confluence integration (if MCP tools for Atlassian are available in the session). Ask: "Would you like me to publish this to Confluence or create a Jira epic?" If yes, use the available MCP tools to do so.
- Confirm completion. Tell the user the file path, how many functional requirements were generated, and how many acceptance criteria scenarios were written.
Interview Mode
When invoked with --interview, conduct a structured interview by asking each question below one at a time. Wait for the user's full answer before proceeding to the next question. Do not ask multiple questions at once.
- Who is the primary user? "Describe the person or system that will use this feature. What is their role, technical level, and what are they trying to accomplish?"
- What problem does this solve? "What is the user currently unable to do, or what pain point does this address? What happens today without this feature?"
- What does success look like? "How will you know this feature is working correctly? What would a user be able to do that they couldn't before?"
- What are the constraints? "Are there technology choices already made (language, framework, cloud provider)? Any performance requirements? Any compliance or regulatory constraints (GDPR, HIPAA, SOC2)?"
- What is explicitly out of scope? "What are you deliberately NOT building in this version? What might people assume is included but isn't?"
- What could go wrong? "What edge cases or error scenarios do you anticipate? What happens if the user provides invalid input? What if a dependency is unavailable?"
- Who else is affected? "Are there other systems, teams, or users impacted by this change? Any backwards compatibility concerns?"
After collecting all answers, generate the full requirements document using the same format as description mode (FR-NNN requirements, NFRs, Gherkin scenarios, etc.), grounded in the interview answers. Include an Interview Summary section at the very top of the generated document (before Overview) that captures each answer in 1–2 sentences.
## Interview Summary
**Primary user:** ...
**Problem being solved:** ...
**Success criteria:** ...
**Constraints:** ...
**Out of scope:** ...
**Risk areas:** ...
**Other affected parties:** ...
Then continue with the standard sections (Overview, Stakeholders, Functional Requirements, etc.) as defined in the Output Format Reference.
Output Format Reference
# Requirements:
**Version**: 1.0
**Date**:
**Status**: Draft
---
## Overview
...
## Stakeholders
| Stakeholder | Role | Primary Need |
...
## Functional Requirements
**FR-001**: The system shall...
**FR-002**: The system shall...
## Non-Functional Requirements
### Performance
...
### Security
...
### Accessibility
...
### Scalability
...
## Acceptance Criteria
```gherkin
Scenario: ...
Given ...
When ...
Then ...
Edge Cases & Error Scenarios
- ...
Out of Scope
- ...
Assumptions
- ...
Open Questions
- ...
## Source & license
This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.
- **Author:** [RealDougEubanks](https://github.com/RealDougEubanks)
- **Source:** [RealDougEubanks/ClaudeMarketplace](https://github.com/RealDougEubanks/ClaudeMarketplace)
- **License:** MIT
Install and usage instructions live in the source repository linked above.
Reviews
No reviews yet, be the first.
Write a review
Versions
- v0.1.0 Imported from the upstream source.