원클릭으로
cc-create-spec
Create Technical Specification
Codex 또는 Claude로 설치 이 Prompt를 복사해 Codex, Claude 또는 다른 어시스턴트에 붙여 넣으면 Skill 페이지를 검토하고 설치를 진행할 수 있습니다.
메뉴
Create Technical Specification
Codex 또는 Claude로 설치 이 Prompt를 복사해 Codex, Claude 또는 다른 어시스턴트에 붙여 넣으면 Skill 페이지를 검토하고 설치를 진행할 수 있습니다.
SOC 직업 분류 기준
| name | cc-create-spec |
| description | Create Technical Specification |
You are tasked with creating comprehensive technical specifications by synthesizing outputs from research and implementation planning into a high-level architectural document.
When this command is invoked:
Check if parameters were provided:
If no parameters provided, respond with:
I'll help you create a technical specification. This spec will synthesize your research findings and implementation plan into a comprehensive architectural document.
Please provide:
1. Path to research document (e.g., `thoughts/shared/research/YYYY-MM-DD-topic.md`)
2. Path to implementation plan (e.g., `thoughts/shared/plans/YYYY-MM-DD-topic.md`)
3. Any additional context or constraints
4. Link to product requirements (Linear ticket, PRD, etc.)
I'll analyze these inputs and create a detailed tech spec that covers problem statement, architecture, implementation approach, and roll-out strategy.
Tip: You can invoke this command with documents directly:
`/create-spec thoughts/shared/research/2025-01-08-feature.md thoughts/shared/plans/2025-01-08-feature.md`
Then wait for the user's input.
Read all mentioned files immediately and FULLY:
Verify understanding with the user:
Based on the research and plan documents, I understand we're building:
**Problem**: [Summary of user problem from research]
**Solution**: [High-level approach from plan]
**Scope**: [Key phases from implementation plan]
Before I create the spec, I need clarification on:
- [Business context question]
- [Stakeholder identification]
- [Success metrics question]
- [Rollout approach question]
Are there any additional constraints or requirements I should consider?
Present a proposed structure to the user:
Here's the spec outline I'll create:
**Table of Contents**
# Problem Statement
[Summarize user problem and customer value]
- Link to PRD
## Open Questions
[Remaining questions the team needs answered]
## Assumptions
[Assumptions being made in this doc]
## Glossary [optional]
[Terms and definitions]
## Requirements
[Functional and non-functional requirements]
## Out of scope
[What we're specifically NOT solving]
## Prior References
[Links to previous tech specs with relevance notes]
# Proposed Solution (Architecture Design)
[Solution description with complexity callouts]
## Alternate Solution(s)
[Other solutions considered and why abandoned]
## Risk and mitigations
[Potential risks and mitigation strategies]
# Roll-out Plan
[How this should be rolled out]
## Permissions / Auditing / Monitoring
[Permissions needed, audit events, monitoring metrics]
# References
[External documents and API docs]
Does this structure capture everything needed?
Determine the filename:
thoughts/shared/specs/YYYY-MM-DD-PRO-XXXX-brief-description.md
YYYY-MM-DD-PRO-XXXX-description.md where:
thoughts/shared/specs/2025-01-15-PRO-1234-scheduled-config-changes.mdthoughts/shared/specs/2025-01-15-role-management.mdUse this template structure:
# [Feature Name] Technical Specification
**Table of Contents**
# Problem Statement
[Summarize the user problem and the customer value we'd like to create. e.g. for summative bonus, it'd be adding policy-level reporting, chargebacks for churned policies. Link to PRD]
Clearly articulate the problem from the user's perspective:
- What pain point are users experiencing?
- What can't users do today that they need to do?
- What inefficiency or risk does this create?
## Open Questions
[List remaining questions the team needs answered]
## Assumptions
[What assumptions are being made in this doc]
## Glossary [optional]
[List terms and their definitions used within the doc or links to our internal glossary]
## Requirements
[]
## Out of scope
[What are you specifically not trying to solve, link to any use cases & examples from PRD]
## Prior References
[Link to previous tech specs. If possible, add something that summarizes and clarifies previous tech specs.
e.g. for tech specs that you're referencing give a one liner of each one's relevance and takeaways and/or add a diagram or table]
# Proposed Solution (Architecture Design)
[Describe your solution and requirements. Add call outs for specific points of complexity]
## Alternate Solution(s)
[What other solutions were considered? Why were they abandoned?]
## Risk and mitigations
[]
# Roll-out Plan
[How should this be rolled out]
## Permissions / Auditing / Monitoring
[What are the permissions we need to support]
[Audit log events]
[What monitoring metrics should be tracked?]
# References
[Other external documents and references e.g. external API docs]
Present the draft spec location:
I've created the technical specification at:
`thoughts/specs/YYYY-MM-DD-PRO-XXXX-description.md`
Please review and let me know:
- Is the problem statement clear?
- Are the requirements complete?
- Are assumptions documented?
- Should any alternate solutions be added?
- Are open questions listed?
- Are risks and mitigations comprehensive?
Iterate based on feedback:
Continue refining until the user is satisfied
Synthesize, Don't Copy:
Business Focus First:
Completeness:
Clarity:
Actionability:
No Placeholders:
Minimize code snippets
This command builds on:
/research-codebase: Provides detailed technical findings and current state analysis/create-plan: Provides implementation approach, phases, and success criteriaThe tech spec synthesizes both into a comprehensive architectural document that:
User: /create-spec thoughts/shared/research/2025-01-15-pay-cycle.md thoughts/shared/plans/2025-01-15-pay-cycle.md
Manage and configure Omarchy Linux systems. Use when user asks about Omarchy, Hyprland, themes, keybindings, system config, or any omarchy-* commands.
Use when building or modifying Datastar-powered UI — SSE streams, templ components, form handlers, signals, morph patterns, loading indicators. Enforces the Tao of Datastar (backend is source of truth, CQRS, fat morph, signals sparingly). Use for any work touching .templ files, SSE handlers, or Datastar attributes.
Create or fix static HTML slide decks powered by Datastar signals. Use when asked to make "datastar slides", "HTML slide deck", "slide deck with Datastar buttons", "arrow-key slides", or fix broken Datastar slide navigation. Covers v1 colon attribute syntax, local UI signals, keyboard handlers, dot indicators, and compact slide layout.
Address GitHub PR review comments in pi. Use when asked to handle PR comments, respond to review feedback, fix review comments, reply to GitHub PR comments, or work through unresolved review threads. Fetches comments with gh, batches triage with the user, applies fixes on the correct branch, and replies on GitHub using pi's /answer flow.
Research Codebase
Build and customize pi itself (not end-user task execution). Use when asked to create or modify pi config, extensions, AGENTS.md behavior, skills/prompts/themes loading, package wiring, or integration surfaces for external tools/services.