About to-spec
to-spec turns conversation context and codebase knowledge into a feature specification, then publishes it to the chosen tracker or local Markdown. It minimizes follow-up questions while confirming testing seams.
Key features
- 01Synthesize context into a specification
- 02Consult repository glossary and ADRs
- 03Resolve a publication destination
- 04Confirm testing seams in one batch
- 05Publish with ready-for-agent status
Use cases
- 01Draft a feature spec from an established discussion
- 02Document implementation and testing decisions
- 03Publish a spec locally or to a requested tracker
Install
Copy the command and run it in a terminal.
npx skills add https:/ / github.com/ c0nant/ skills --skill to-specAfter installing, type/to-specin Claude Code to confirm it loaded.
Bundled files
References and scripts installed along with SKILL.md.
RESUMO.mdTRADUCAO.pt-BR.mdagents/openai.yamlDetails
From the project READMEThis skill takes the current conversation context and codebase understanding and produces a spec. Default to synthesizing what you already know: by the time this skill runs, the deciding is usually done, so do not reopen it as an interview.
You may ask questions when you genuinely need to. The bar is a gap you cannot close from the conversation, the codebase, GLOSSARY.md, or the ADRs, and that would otherwise force you to invent a decision the user never made. When you hit that bar:
- Ask in one batch, before writing the spec, not question by question.
- Keep it to the few that actually change what the spec says.
- Carry your best guess with each one, so the user can answer by confirming.
- The seams (step 3) are one of those questions: put your proposed seams in the batch as the best guess. If nothing else clears the bar, ask one short question about the seams only.
- Then write the spec. Never stall waiting on answers you could have assumed and flagged.
If nothing clears that bar, write the spec without asking anything, except the seams: those are always confirmed, in the batch or as the one short seams-only question.
Process
-
Resolve where to publish. This one never needs a question: the chain below always lands somewhere, so pick a destination silently and carry it into step 4 rather than presenting a picker.
- If the user explicitly asked this run to publish to GitHub, GitLab, local markdown, or another tracker: use that.
- Else if
docs/agents/issue-tracker.mdexists (written by/setup-skills) and names a tracker: use that. - Else: Local markdown (
.scratch/<feature-slug>/SPEC.md). Do not probegit remote. Do not present a destination picker.
For GitHub or GitLab, follow the conventions and triage-label mappings in
docs/agents/issue-tracker.mdwhen present; otherwise use<destination-conventions>below. Run/setup-skillsto configure those conventions and a triage-label vocabulary. -
Explore the repo to understand the current state of the codebase, if you haven't already. Use the project's domain glossary vocabulary throughout the spec, and respect any ADRs in the area you're touching.
-
Sketch out the seams at which you're going to test the feature. Existing seams should be preferred to new ones. Use the highest seam possible. If new seams are needed, propose them at the highest point you can. The fewer seams across the codebase, the better - the ideal number is one.
Do not confirm them separately: they go to the user in the single batch described above, and the agreed seams are written into the spec's Testing Decisions.
-
Write the spec using the template below, then publish it to the destination resolved in step 1. Apply the
ready-for-agenttriage label: no need for additional triage. (For Local markdown, "applying a label" means writing aStatus: ready-for-agentline near the top of the file instead.)
- GitHub:
gh issue create --title "..." --body "..."(heredoc for the multi-line body). Triage labels via--label. - GitLab:
glab issue create --title "..." --description "..."(heredoc for the multi-line description). Triage labels via--label. - Local markdown: write
.scratch/<feature-slug>/SPEC.md, creating the directory if needed. Record triage state as aStatus:line near the top of the file instead of a label.
Problem Statement
The problem that the user is facing, from the user's perspective.
Solution
The solution to the problem, from the user's perspective.
User Stories
A LONG, numbered list of user stories. Each user story should be in the format of:
- As an , I want a , so that
This list of user stories should be extremely extensive and cover all aspects of the feature.
Implementation Decisions
A list of implementation decisions that were made. This can include:
- The modules that will be built/modified
- The interfaces of those modules that will be modified
- Technical clarifications from the developer
- Architectural decisions
- Schema changes
- API contracts
- Specific interactions
Do NOT include specific file paths or code snippets. They may end up being outdated very quickly.
Exception: if a prototype produced a snippet that encodes a decision more precisely than prose can (state machine, reducer, schema, type shape), inline it within the relevant decision and note briefly that it came from a prototype. Trim to the decision-rich parts, not a working demo, just the important bits.
Testing Decisions
A list of testing decisions that were made. Include:
- A description of what makes a good test (only test external behavior, not implementation details)
- The agreed seams, one per bullet, so
/to-ticketsand/tddcan pick them up - Which modules will be tested
- Prior art for the tests (i.e. similar types of tests in the codebase)
Out of Scope
A description of the things that are out of scope for this spec.
Further Notes
Any further notes about the feature.