
Writing Specs
Applies a shared quality bar so specs stay short, honest, and high-signal
- Repository
- JetBrains/thinkrail
- Stars
- 7,089
- Licence
- Apache-2.0
- Files
- 1 file
What it does
This is a style and quality guide that tells an AI agent how to write project specification documents inside a structured workflow. When drafting or revising a goal, architecture, or module spec, the agent keeps specs short, marks unconfirmed inferences, avoids placeholder sections, keeps rationale in the spec rather than in code comments, and avoids listing versions or phases in product goal docs. It does not run any workflow or code itself.
Great for
- Developers using the ThinkRail workflow who want consistent, concise specs
- Teams that keep product goal documents and module specs in a repository and want an agent to follow shared rules for them
- Anyone building their own spec-writing skills who wants an example of a clear quality bar
Skip it if
- Users who need an agent to execute tasks or run a workflow, because this skill contains only guidance and nothing to execute (the skill says so itself)
- Projects without the ThinkRail spec-graph conventions, since it refers to `spec_*` tools and `status` fields that other setups may not have
Try asking
- “Draft a goal-and-requirements doc for my new CLI tool, keeping it free of MVP or phase lists”
- “Revise this module spec so it is shorter and marks anything you inferred as unconfirmed”
- “Move the rationale comments in this module into its owning spec and leave a one-line pointer”
Install in 3 steps
- 1
Download the skill
The ZIP contains the 1 original files, the Apache-2.0 licence and an attribution note.
writing-specs.zip - 2
Add it to your agent
Claude Code (all projects)
mkdir -p ~/.claude/skills && unzip writing-specs.zip -d ~/.claude/skills/Claude Code / Agent SDK (this project only)
mkdir -p .claude/skills && unzip writing-specs.zip -d .claude/skills/Claude apps (web/desktop): Settings → Capabilities → Skills → Upload skill, and choose the ZIP.
- 3
Ask for the task
No special command needed — the agent reads the skill's description and loads it when your request matches. Try one of the prompts above.
Get better results
- Use it alongside the ThinkRail workflow skills that reference it at spec-drafting steps, since it sets the bar rather than the process.
- Review drafted specs yourself, because the skill makes the user, not the agent, responsible for moving a spec out of draft status.
- Pair it with the spec-graph skill, which covers the frontmatter and linking mechanics this skill depends on.
What's inside
- SKILL.md
Redistributed unmodified from JetBrains/thinkrail under the Apache-2.0 licence. The assessment was drafted with AI from the skill's files on 2026-11-02. Skills can instruct your agent to run code — read SKILL.md and any scripts before installing.