Reshape my PREVIOUS message using extra instructions, then save it as a slash command
Scanned 9/6/2026
Install to Claude Code
npx -y skills add fayis672/slash-IT --skill create-cmd-instruct --agent claude-codeInstalls into .claude/skills of the current project.
Are you the author of Create Cmd Instruct?
Add the live security badge to your README — it updates automatically with every re-scan.
[](https://www.skillsdirectory.com/skills/fayis672-create-cmd-instruct)More formats (shields.io, HTML) on the badges page.
---
description: Reshape my PREVIOUS message using extra instructions, then save it as a slash command
argument-hint: <command-name> | <how to change it> [--global | --project]
disable-model-invocation: true
---
The user wants to take their **previous message** — the most recent user turn in this
conversation *before* this invocation — **modify it according to extra instructions they provide
here**, and save the result as a reusable custom slash command.
Raw arguments: `$ARGUMENTS`
The arguments contain a command **name** and **instructions** for how to change the prompt,
separated by a `|` pipe. An optional scope flag may appear in the name part. Example:
`/slash-it:create-cmd-instruct daily-report --project | output a markdown table and ask for the date`
Do this:
1. Split `$ARGUMENTS` on the first `|`.
- Left side = the command name (and optional scope flag).
- Right side = the user's instructions for reshaping the prompt.
If there is no `|`, treat the first token as the name and the rest as instructions.
2. From the left side, extract the **name** and any **scope flag**:
- `--global` / `-g` / `global` -> system-wide. `--project` / `-p` / `project` -> this project.
3. If the name is empty, STOP and ask for it (required). If the instructions are empty, ask what
changes the user wants.
4. **Determine where to save (scope):**
- If a scope flag was given, use it.
- Otherwise ASK with AskUserQuestion: "Where should `/<name>` be saved?" -> **Global** or
**Project**.
- Global path: `~/.claude/commands/`
- Project path: `<current working directory>/.claude/commands/` — create if missing; fall back
to Global if not in a project and say so.
5. Normalize the name: lowercase, kebab-case, strip any leading `/` and `.md`.
6. Find the user's previous prompt (the message right before this call).
7. Produce a new prompt = the previous prompt **transformed according to the instructions**, plus
light optimisation (clarity, structure, grammar) while honoring the original intent. Parameterize
with `$ARGUMENTS`/`$1`/`$2` where it makes the command reusable.
8. Briefly show the user the resulting prompt so they can see what was saved.
9. Add YAML frontmatter with a one-line `description` and an `argument-hint` if parameters exist.
10. Write the file to `<chosen path>/<name>.md`.
11. If that file already exists, show its current contents and ask before overwriting.
12. Confirm: print the saved file path, the scope, and that `/<name>` is now available.
Is this your skill, or is something wrong with this listing? Request removal or report an issue. Author removals are honored within 72 hours.
No comments yet. Be the first to comment!