Skip to content
Agentmart
Submit a tool

sdd:auto

SDD — Spec-Driven Development: run the full pipeline automatically.

Skill by alfredoperez

IndexedMIT★ 2
Write a review

Install

Works with Claude Code, Claude apps, Codex, Cursor, Gemini CLI, GitHub Copilot, VS Code, OpenCode, Goose, Windsurf. Read the repository before you give any tool access to your data.

npx agentmarket install skill/alfredoperez-sdd-auto

Asks which of your tools to add it to (Claude Code, Codex, Cursor, VS Code, Windsurf, Gemini CLI, Claude Desktop) and shows the change first. See https://agentmart.dev/cli.

Get the weekly diff

The weekly diff: one email a week with new agent tools, notable updates and grade changes, plus security alerts for tools you follow. Double opt-in, one-click unsubscribe.

We use your email only for this newsletter. Privacy

About

SDD — Spec-Driven Development: run the full pipeline automatically.

Use cases: data

No reviews yet. Installed it? Reviews come from verified installers only. How reviews work

Permissions

What this skill can do and what it asks for, from its SKILL.md. Checked Oct 25, 2026.

Tools
No tools found
Secrets
None asked for
Install scripts
None found

Network hosts it talks to 1

  • codeload.github.com · found in install command

Files it touches

  • ~/.claude/skills/sdd-auto

What's inside

Skill name
sdd:auto
Description
SDD — Spec-Driven Development: run the full pipeline automatically.
License
MIT
Instructions
71 lines in SKILL.md · view on GitHub

Agent Skills spec issues

  • name: only lowercase letters, digits and single hyphens allowed
  • name: "sdd:auto" does not match folder "auto"

Files (1)

  • SKILL.md

SKILL.md (first lines)

## Shared Instructions

- [Transition Logging](../../lib/instructions/transition-logging.md) — append a transition entry on every `.spec-context.json` write

## User Input

```text
$ARGUMENTS
```

If `$ARGUMENTS` is empty, stop and say: "Provide a feature description: `/sdd:auto <description>`"

---

## Steps

### 1. Run Specify, Mark Auto, and Gate

**CRITICAL: This entire step is one continuous action — your turn is not complete until either the AskUserQuestion (sub-actions 5 + 6, normal mode) has been called, or you have entered the auto-advance loop (Step 2, minimal mode). The sub-actions below are NOT separate stop points.**

1. Invoke `/sdd:specify` via the **Skill** tool, passing `$ARGUMENTS` as the argument.

2. When specify completes, find the most recently modified directory under `specs/` that contains a `.spec-context.json`. Record its slug (`{NNN}-{slug}`).

3. Read `specs/{NNN}-{slug}/.spec-context.json` and set `auto` to `true`. Write the updated file back, preserving all existing fields and appending a transition entry per [transition-logging](../../lib/instructions/transition-logging.md). Do not set `auto` earlier — specify's footer behavior depends on `auto` being `false` during its run.

4. Determine complexity from the file you just read:
   - `currentStep: "tasks"` → **minimal** (specify already wrote plan.md + tasks.md). Proceed directly to Step 2 (Auto-Advance Loop). No gate.
   - `currentStep: "specify"` → **normal**. Continue to sub-actions 5 + 6 in this same step.

5. **Normal mode only** — read `specs/{NNN}-{slug}/spec.md` and display exactly:

   ```
   📋 **Spec review — approval needed**

   {Feature Name} (`{NNN}-{slug}`)

   {1-line summary from the spec's Summary section}
   {N} requirements · {N} scenarios

   📂 `specs/{NNN}-{slug}/spec.md`
   ```

6. **Normal mode only** — call **AskUserQuestion** with these options:
   - **Continue** — proceed with auto-advance through remaining phases
   - **Edit spec** — user provides edit notes in the "Other" field; apply changes to spec.md, then re-do sub-actions 5 + 6

   Do not proceed past sub-action 6 until the user approves.

---

### 2. Auto-Advance Loop

**CRITICAL: This is a loop. Do NOT stop or end your response after a Skill tool invocation returns. You MUST read `.spec-context.json` and continue the loop until a termination condition is met.**

Loop until complete:

1. Read `specs/{NNN}-{slug}/.spec-context.json`. If `paused` is `true`, display: "⏸ Spec is paused. Run `/sdd:resume` to continue." and stop.
2. Invoke `/sdd:resume {NNN}-{slug}` via the **Skill** tool
3. **After the Skill tool returns, you MUST continue — do NOT stop.** Read `specs/{NNN}-{slug}/.spec-context.json`
4. If `next` is `"done"`, set `auto` to `false` in .spec-context.json (append a transition entry per [transition-logging](../../lib/instructions/transition-logging.md)), then stop — the pipeline is complete
5. If `currentStep` is `"implement"` and `progress` is `null` and `next` is `"done"`, set `auto` to `false` in .spec-context.json (append a transition entry per [transition-logging](../../lib/instructions/transition-logging.md)), then stop — shipped
6. **Otherwise, loop back to step 1. Do NOT stop.**

**Notes:**
- CP1 (Code Review) is handled by the implement skill's own AskUserQuestion — it will pause for user approval automatically. This skill does not bypass it.
- If any skill stops with a blocker (architectural change, impossible task), the loop naturally stops because the skill will have asked the user a question.
- The loop reads .spec-context.json after each invocation to determine if the pipeline is done, rather than counting steps.
- The only valid reasons to stop are: `next` is `"done"`, spec is paused, or a skill asked the user a question.

From alfredoperez/sdd, MIT.

More from alfredoperez/sdd

Signals

Stars
2
Forks
0
Open issues
0
Last commit (this folder)
Apr 27, 2026 (6 months ago)

From GitHub (fetched Oct 24, 2026).

Details

License
MIT
First seen
Oct 24, 2026
Last checked
Oct 24, 2026

More from alfredoperez

Cite this page

sdd:auto skill. Agent Mart. https://agentmart.dev/skills/alfredoperez-sdd-auto (last updated 2026-10-24)