Skip to main content
skill CX Codex

spark-recipe-draft-batch — Claude Code skill: batch draft review

A skill that walks every unfinished email draft, reads the thread behind it, and decides what to edit, send, or trash.

Install into Claude Code · Skills CLI (Vercel) All methods and platforms ↓
npx skills add readdle/spark-cli-skills --skill spark-recipe-draft-batch -a claude-code
npx skills add readdle/spark-cli-skills --skill recipe-draft-batch -a claude-code
Open on GitHub

What spark-recipe-draft-batch does

Drafts pile up when you start a reply and never finish it. This skill clears the backlog in one pass: it lists everything with spark emails Drafts, then opens each thread with spark thread <id> to see whether someone already replied or the topic is closed.

Each draft gets one of three outcomes. Edit it in place with spark draft --edit <id> --body "..." or a new subject line. Keep it as is, and the skill just flags it for you. Discard it with spark action moveToTrash <id> — that one is reversible, the draft comes back from trash. The run ends with a summary of how many were reviewed, edited, left ready, and discarded.

It needs the triage access level and the base use-spark skill, which holds the command reference and filter syntax. Without that, the recipe steps have nothing to lean on. The author suggests running it weekly so the pile never grows.

What it helps with

  • clear a backlog of unfinished email drafts in one pass
  • check whether a draft went stale after a new reply landed
  • fix a draft's body or subject without recreating it
  • move dead drafts to trash
  • get a count of edits, ready-to-send, and discarded drafts

Example prompts for the agent

  • list every draft in Drafts and walk me through them one by one
  • read the thread for each draft and tell me if it's still relevant
  • clean up my drafts: what to send, what to edit, what to trash

How to install spark-recipe-draft-batch

1Skills CLI (Vercel)

npx skills add readdle/spark-cli-skills --skill spark-recipe-draft-batch -a claude-code
npx skills add readdle/spark-cli-skills --skill recipe-draft-batch -a claude-code

2Manually: put the files in place

~/.claude/skills/recipe-draft-batch/SKILL.md

Copy the whole skill folder; for one project use .claude/skills/recipe-draft-batch/

Step by step, with commands for each platform: How to install skills in Claude Code: five ways

Questions about spark-recipe-draft-batch

How do I install spark-recipe-draft-batch in Claude?

Run npx skills add readdle/spark-cli-skills --skill spark-recipe-draft-batch -a claude-code. The skill file then sits at ~/.claude/skills/recipe-draft-batch/SKILL.md.

What does spark-recipe-draft-batch do?

It runs a batch review of pending drafts: reads each thread for context, edits the draft, marks it ready to send, or trashes it, then reports the totals.

What does the skill require to work?

The triage access level and the base use-spark skill for command reference and filter syntax. Note that draft --edit updates the existing draft rather than creating a new one.

Source text

SKILL.md · MIT
---
name: spark-recipe-draft-batch
description: >-
  Review and finalize multiple pending drafts in a batch session: read
  context, edit, or discard.
metadata:
  version: 1.0.0
  requires:
    skills:
      - use-spark
    accessLevel: triage
---

# Recipe: Draft Batch Review

Review and finalize multiple pending drafts in a single session. Read context for each, decide whether to edit, keep, or discard.

**Prerequisite:** Read the `use-spark` base skill for command reference and filter syntax.

**Access level required:** triage.

## Steps

### Step 1: List all pending drafts

```bash
spark emails Drafts
```

Review the list of drafts - note subjects, recipients, and dates.

### Step 2: Review each draft in context

For each draft, read the full thread to understand the context:

```bash
spark thread <id>
```

Check if the conversation has moved on since the draft was created - someone else may have replied, or the topic may be resolved.

### Step 3: Decide on each draft

For each draft, present the options to the user:

**Edit** - the draft needs changes before sending:

```bash
spark draft --edit <id> --body "Updated content..."
spark draft --edit <id> --subject "Updated subject"
```

**Keep as-is** - the draft is ready to send. Note it for the user.

**Discard** - the draft is no longer needed. Move to trash:

```bash
spark action moveToTrash <id>
```

### Step 4: Report

Summarize: "Reviewed N drafts. M edited, K ready to send, L discarded."

## Tips

- Drafts accumulate when the user starts replies but doesn't finish them. This recipe clears the backlog.
- Always read the thread before editing a draft - the context may have changed since it was created.
- If a thread has new replies since the draft was created, the draft may be outdated and need rewriting.
- `draft --edit` updates the draft in place - it doesn't create a new one.
- For drafts that are replies (`--reply-to`), read the original thread to verify the reply still makes sense.
- Run this recipe weekly to prevent draft accumulation.
- Discarding via `moveToTrash` is reversible - the draft can be recovered from trash if needed.

© readdle · MIT · readdle/spark-cli-skills

Similar skills

Everything for Notes, tasks and projects →

From the same set19

Project sponsors

Companies that keep this analytics open
Gold sponsor

Impressions, clicks and code copies — open statistics. We keep the numbers public for the same reason NeBlask itself is public. Become a sponsor