How do I install the Write Spec skill?

Run npx skillmds@latest add alaimo-labs/write-spec in your terminal (requires Node.js), paste this page's agent-chat prompt into Claude, Cursor, or any MCP-connected agent, or download the SKILL.md file and copy it into your agent's skills directory.

Is Write Spec safe to use?

This skill has not completed SkillMD's automated safety review yet. Capability flags: docs only. SkillMD never runs a skill's scripts for you; review the SKILL.md before installing.

Which AI agents work with Write Spec?

This skill is tagged as working with Claude Code, Claude.ai, OpenAI Codex. SKILL.md is an open format, so most agents that read a skills directory can load it too.

Is Write Spec free to use?

Yes. Installing skills from SkillMD is free, and the skill stays under its author's original license.

Who published Write Spec?

alaimo-labs (@alaimo-labs) published this skill. Their other Agent Skills are listed on their SkillMD profile.

Write Spec by alaimo-labs — Agent Skill

Name: Write Spec
Rating: 4.2 (5 reviews)
Author: alaimo-labs

File contents

/write-spec

Draft a feature spec from existing discovery artifacts, following the feature-specs skill.

Input: $ARGUMENTS

Workflow

Resolve the bet. If the arguments name a feature idea or an insight file, start there. Otherwise read the recent files in product/insights/, propose the insight that looks highest-impact as the one to bet on, and confirm with the user before proceeding.
Load the evidence. Read product/overview.md (belief registry included — beliefs already registered for this feature or the product bear on the bet), the personas in product/personas/ relevant to the chosen insight, the interview transcripts the insight cites, and any market research in product/research/ that bears on the bet. Identify the primary persona for the journey and confirm the choice if it isn't obvious.
Draft per the feature-specs skill: problem with cited evidence, user journey from the primary persona's point of view, 3–5 critical user stories with observable acceptance criteria, the falsifiable hypothesis block, and assumptions. Where the evidence is thin, the claim becomes an assumption — never invent quotes or journey steps.
Present and iterate. Show the draft; walk the user through the journey and the hypothesis in particular (those carry the most judgment). Adjust until they own it — the spec is their bet, not the agent's.
Register the assumptions in the overview — the spec keeps no list of its own. product/overview.md is the single belief registry. Reuse the feature slug from the idea brief when the spec descends from one; otherwise use the spec's slug. For each assumption: already registered → the spec's Assumptions section references it; new → propose appending [feature: {slug}] [risk] {assumption} to the overview's unverified beliefs — risk classified per the feature-specs skill's four categories, with their default owners — and write nothing without the user's approval.
Save to product/specs/{YYYY-MM-DD-HHMM}-{slug}.md and suggest the natural next step in one line: run the critique panel over it (/critique-spec).

---name: write-specdescription: Draft an evidence-grounded feature spec from your insights and personas — problem, user journey, critical user stories with acceptance criteria, falsifiable hypothesis---# /write-specDraft a feature spec from existing discovery artifacts, following the `feature-specs` skill.Input: $ARGUMENTS## Workflow1. **Resolve the bet.** If the arguments name a feature idea or an insight file, start there. Otherwise read the recent files in `product/insights/`, propose the insight that looks highest-impact as the one to bet on, and confirm with the user before proceeding.2. **Load the evidence.** Read `product/overview.md` (belief registry included — beliefs already registered for this feature or the product bear on the bet), the personas in `product/personas/` relevant to the chosen insight, the interview transcripts the insight cites, and any market research in `product/research/` that bears on the bet. Identify the primary persona for the journey and confirm the choice if it isn't obvious.3. **Draft** per the `feature-specs` skill: problem with cited evidence, user journey from the primary persona's point of view, 3–5 critical user stories with observable acceptance criteria, the falsifiable hypothesis block, and assumptions. Where the evidence is thin, the claim becomes an assumption — never invent quotes or journey steps.4. **Present and iterate.** Show the draft; walk the user through the journey and the hypothesis in particular (those carry the most judgment). Adjust until they own it — the spec is their bet, not the agent's.5. **Register the assumptions in the overview — the spec keeps no list of its own.** `product/overview.md` is the single belief registry. Reuse the feature slug from the idea brief when the spec descends from one; otherwise use the spec's slug. For each assumption: already registered → the spec's Assumptions section references it; new → propose appending `[feature: {slug}] [risk] {assumption}` to the overview's unverified beliefs — risk classified per the `feature-specs` skill's four categories, with their default owners — and write nothing without the user's approval.6. **Save** to `product/specs/{YYYY-MM-DD-HHMM}-{slug}.md` and suggest the natural next step in one line: run the critique panel over it (`/critique-spec`).

Frequently asked questions

Write Spec