New: 98,000+ skills indexed from GitHub repos with 100+ stars · every skill safety-checkedBrowse top skillsTop skills
Esc to closeBrowse all skills
Submit a skillBrowse skills
Submit a skillBrowse skills

Agent skill

Prompt

by chrisblattman in chrisblattman/claudeblattman

Turn a rough request into a clear, structured prompt, show it, then run it.

MITAuto-check passed

Install Prompt

skills CLI
$ npx skills add chrisblattman/claudeblattman --skill prompt -a claude-code

Project install by default; add -g for ~/.claude/skills/.

GitHub CLI
$ gh skill install chrisblattman/claudeblattman prompt --agent claude-code

Project scope by default; add --scope user for a personal install. Needs GitHub CLI 2.90.0 or later (public preview).

Manual copy
$ git clone --depth 1 https://github.com/chrisblattman/claudeblattman.git skills-src && mkdir -p .claude/skills && cp -r skills-src/plugins/starter-kit/skills/prompt .claude/skills/prompt && rm -rf skills-src

Use ~/.claude/skills/ instead of .claude/skills for a personal install. The folder must contain SKILL.md.

Claude Code skills documentation · loads skills from .claude/skills/

Facts

Skill name
prompt
GitHub stars
463
Token cost
~1.2k tokens
SKILL.md length
587 words
Files
3 (incl. references)
Skills in repo
10
Repo updated
First seen
Licence
MIT

At a glance

Turn a rough request into a clear, structured prompt, show it, then run it.

  • Works in 9 steps: Parse the intent. Extract the core task,… → Calibrate depth using the heuristic in… → Format the request into a structured… → …
  • Your request is rough
  • SKILL.md covers Reference files (read before…, Input, Instructions and Important
  • Instructions only: no scripts, shell commands, URLs or credentials in SKILL.md

What it does

Prompt is an agent skill from chrisblattman/claudeblattman. Turn a rough request into a clear, structured prompt, show it, then run it. Use when your request is rough, dictated, or messy, when one real ask is buried in a long brain dump, or when you want to refine or improve an existing prompt (refine mode returns the improved prompt without running it).

Its SKILL.md is about 1.2k tokens, which your agent loads only when the skill is triggered. The skill folder holds 3 other files, including reference files (for example `references/formatting-core.md` and `references/refine-mode.md`).

The repository describes itself as: Claude Code for academics — skills, agents, and setup guides. The licence is MIT.

When your agent uses it

  • Your request is rough
  • One real ask is buried in a long brain dump
  • You want to refine
  • Improve an existing prompt (refine mode returns the improved prompt without running it)

Example prompts

  • “/prompt”

Workflow steps

9 steps, taken from the first numbered list in SKILL.md.

  1. Parse the intent. Extract the core task, audience, and desired output from the informal input.
  2. Calibrate depth using the heuristic in formatting-core.md
  3. Format the request into a structured prompt using the elements in formatting-core.md. Apply elements as appropriate — match formatting…
  4. Inject depth directives if Standard or Deep (templates in formatting-core.md). For Light, skip this step entirely.
  5. Show the formatted prompt in a fenced code block so the user can see exactly what will run.
  6. Tool-routing check: if a dedicated research tool or another app would serve this task better (see formatting-core.md), add a one-line note…
  7. Council opt-in: if the input contains the literal token council, do NOT execute directly. After formatting, invoke /council (it ships with…
  8. Execute the prompt immediately — respond to it as if the user had typed it directly (unless step 7's council token was present).
  9. Ask ONE clarifying question ONLY if the ambiguity would lead to a significantly different output. Otherwise, make reasonable assumptions…

What it can do on your machine

Read from SKILL.md and the folder at commit 12e14d4. It shows what the files ask for, not the result of running them.

  • Tool permissions

    Pre-approves nothing: there is no allowed-tools line, so your agent's usual permission prompts apply.

    From allowed-tools in the SKILL.md frontmatter.

  • Runs code

    No scripts in the folder and no shell commands in SKILL.md.

    From the folder's file list and the shell code blocks in SKILL.md.

  • Network

    No URLs in SKILL.md.

    From URLs in SKILL.md, links to its own repository left out.

  • Credentials

    Names no API keys, tokens, secrets or passwords.

    From names ending in _API_KEY, _TOKEN, _SECRET, _KEY or _PASSWORD in SKILL.md.

Context cost

Prompt loads about 1.2k tokens when it runs, and up to ~3.6k if it reads all its reference files. Until then it costs about 76 tokens; SKILL.md has 587 words of instructions outside code blocks.

Always · name and description, kept in context so the agent knows when to use it
~76
When it runs · the whole SKILL.md, loaded when a task matches
~1.2k
With references · SKILL.md plus every file in references/, read only if the agent opens them
~3.6k

Estimates: characters ÷ 4, the usual rule of thumb; real counts depend on the model's tokenizer. Scripts and assets cost tokens only if the agent reads them.

Safety

Auto-check passed

The automated check found no risky patterns in SKILL.md.

Automated static check — not a guarantee. Review scripts before installing. It scans the text of SKILL.md for risky patterns (piping downloads into a shell, reading credential files, hidden Unicode, destructive commands); files beside SKILL.md are not scanned.

SKILL.md

The full file from chrisblattman/claudeblattman at commit 12e14d4, republished under its MIT licence (© chrisblattman). 587 words, ~1,173 tokens.

Download SKILL.mdSave it as .claude/skills/prompt/SKILL.md (or your agent's skills folder). This skill also uses 2 other files; get the full folder from GitHub.
name
prompt
description
Turn a rough request into a clear, structured prompt, show it, then run it. Use when your request is rough, dictated, or messy, when one real ask is buried in a long brain dump, or when you want to refine or improve an existing prompt (refine mode returns the improved prompt without running it).
argument-hint
request

/prompt — Format and Execute

v0.1.0 — adapted from a personal workflow

Take an informal request, format it into a clear structured prompt, show the prompt, then execute it.

Refine mode: If the input starts with refine or refine:, or the user is asking to audit or improve an EXISTING prompt rather than run a new request, read ${CLAUDE_PLUGIN_ROOT}/skills/prompt/references/refine-mode.md and follow it instead of the steps below (audit + improved prompt as output; do not execute).

Reference files (read before formatting)

Read ${CLAUDE_PLUGIN_ROOT}/references/prompt-preferences-TEMPLATE.md — or the user's personalized copy at ~/.claude/starter-kit/my-preferences.md if it exists — plus ${CLAUDE_PLUGIN_ROOT}/references/prompting-guide.md and the skill-local references:

  • ${CLAUDE_PLUGIN_ROOT}/skills/prompt/references/formatting-core.md — formatting elements, depth calibration, depth-injection templates, tool routing
  • ${CLAUDE_PLUGIN_ROOT}/skills/prompt/references/refine-mode.md — only when refine mode triggers

Authority: the user's personal preferences file wins over the general guide wherever they conflict. If no personal copy exists, treat the template's defaults as the user's preferences.

Input

$ARGUMENTS

Instructions

You are a prompt formatter. The user has given you an informal, conversational request (often dictated or roughly typed). Your job:

  1. Parse the intent. Extract the core task, audience, and desired output from the informal input.

  2. Calibrate depth using the heuristic in formatting-core.md:

    • Light (default): format only. No extra directives.
    • Standard: format + append an assumptions/rationale block.
    • Deep: format + append a research/compare/verify block.
    • The user can override with depth:light, depth:standard, or depth:deep.

  3. Format the request into a structured prompt using the elements in formatting-core.md. Apply elements as appropriate — match formatting complexity to task complexity.

  4. Inject depth directives if Standard or Deep (templates in formatting-core.md). For Light, skip this step entirely.

  5. Show the formatted prompt in a fenced code block so the user can see exactly what will run.

  6. Tool-routing check: if a dedicated research tool or another app would serve this task better (see formatting-core.md), add a one-line note before executing. Don't block — just flag it.

  7. Council opt-in: if the input contains the literal token council, do NOT execute directly. After formatting, invoke /council (it ships with this kit) with the formatted prompt as the topic. The token is opt-in only — /prompt never wraps a request in a council on its own. This prevents accidental council dispatches from casual uses.

  8. Execute the prompt immediately — respond to it as if the user had typed it directly (unless step 7's council token was present).

  9. Ask ONE clarifying question ONLY if the ambiguity would lead to a significantly different output. Otherwise, make reasonable assumptions and proceed.

Show full SKILL.md (185 more words)Show less

Important

  • Do NOT over-engineer simple requests. A 1-sentence ask doesn't need a 20-line prompt.
  • When writing a prompt, ask for supporting evidence, criteria, or a brief rationale for choices — never instruct a model to output, echo, or explain its internal chain-of-thought. "Think carefully before responding" (for the model's own benefit) is fine.
  • Light depth is the default — most requests should pass through with formatting only.
  • If the user says "hold," "don't run," or "just format," show the formatted prompt but do not execute it.
  • If executing the formatted prompt would send anything outside this machine (an email, a message, a post, a file share), show the draft and get the user's explicit approval before sending — every time.
  • council token recap: /prompt X depth:deep council → format, then hand to /council. /prompt X → format + execute directly (no council).
  • Use available tools (file access, search, connected apps) when executing if the task requires them.
  • If you used the template because no personal preferences file exists, you may end with one short line: "Tip: run /kit-setup to create your personal preferences file at ~/.claude/starter-kit/my-preferences.md." At most once per session.

© chrisblattman, MIT. Rendered from Markdown: HTML in the file is shown as text, images as links, and headings moved down two levels. Raw file

Files

SKILL.md and 2 other files (references) in plugins/starter-kit/skills/prompt of chrisblattman/claudeblattman.

  • SKILL.md
  • references/formatting-core.md
  • references/refine-mode.md

Open the folder on GitHubat commit 12e14d4

Compare with similar skills

Prompt next to the 5 skills that share the most tags, products or categories with it. Stars are the repository's; “used in” counts other GitHub owners with a copy.

Prompt compared with similar skills
SkillStarsUsed inTokensAuto-checkLicenceRepo updated
Prompt this skillchrisblattman/claudeblattman463—~1.2kAutomated safety check: PassMIT
Ccb ClearSeemSeam/claude_codex_bridge3.6k—~354Automated safety check: PassCustom licence
Ccb ClearSeemSeam/claude_codex_bridge3.6k—~151Automated safety check: PassCustom licence
Ccb ClearSeemSeam/claude_codex_bridge3.6k—~381Automated safety check: PassCustom licence
Ccb ClearSeemSeam/claude_codex_bridge3.6k—~160Automated safety check: PassCustom licence
ShowZimoLiao/scholaraio577—~356Automated safety check: PassMIT

Similar skills

  • Ccb Clear

    SeemSeam/claude_codex_bridge

    Clear CCB managed agent conversation context with ccb clear.

    3.6k GitHub stars~354 tokensUpdated 2 days ago
    Agent WorkflowsAuto-check passed

  • Ccb Clear

    SeemSeam/claude_codex_bridge

    Clear CCB managed agent conversation context with ccb clear.

    3.6k GitHub stars~151 tokensUpdated 2 days ago
    Agent WorkflowsAuto-check passed

  • Ccb Clear

    SeemSeam/claude_codex_bridge

    Clear conversation context for one or more mounted CCB agents using ccb clear.

    3.6k GitHub stars~381 tokensUpdated 2 days ago
    Agent WorkflowsAuto-check passed

  • Ccb Clear

    SeemSeam/claude_codex_bridge

    Clear one or more CCB managed agent contexts with ccb clear without deleting project state.

    3.6k GitHub stars~160 tokensUpdated 2 days ago
    Agent WorkflowsAuto-check passed

  • Show

    ZimoLiao/scholaraio

    A skill your agent uses when the user wants to read a paper or inspect metadata, abstract, conclusion, full text, figures, or detail levels L1 through L4.

    577 GitHub stars~356 tokensUpdated 14 days ago
    Auto-check passed

  • Writing Clearly And Concisely

    JanDeDobbeleer/oh-my-posh

    Apply whenever writing content a human will read — emails, documents, reports, documentation, messages, UI copy, commit messages, explanations, or any other prose.

    24k GitHub stars~1k tokensUpdated yesterday
    DevelopmentAuto-check passed

More from chrisblattman/claudeblattman

All 10 skills in this repo

  • Ask Agents

    chrisblattman/claudeblattman

    Send a task or question to Claude Code and/or Gemini as external subagents and report back what they said.

    463 GitHub stars~1.1k tokensUpdated 2 days ago
    Auto-check passed

  • Council

    chrisblattman/claudeblattman

    Dispatch a small panel of critic agents in parallel on a plan, draft, or decision, then synthesize their raw critiques in a separate fresh-context pass.

    463 GitHub stars~2.1k tokensUpdated 2 days ago
    Auto-check passed

  • Prompt

    chrisblattman/claudeblattman

    Turn a rough, dictated, or ambiguous request into a clear, structured task, then carry it out.

    463 GitHub stars~241 tokensUpdated 2 days ago
    Auto-check passed

  • Review Plan

    chrisblattman/claudeblattman

    Review a proposed plan with a single-pass structured critique and a clear verdict.

    463 GitHub stars~249 tokensUpdated 2 days ago
    Auto-check passed

  • Kit Setup

    chrisblattman/claudeblattman

    One-time starter-kit setup: detects optional AI engines (Codex, Gemini), confirms each with you, writes the engines config, and optionally personalizes how Claude writes for you.

    463 GitHub stars~1.2k tokensUpdated 2 days ago
    Auto-check passed

  • Review Plan

    chrisblattman/claudeblattman

    Stress-test a plan before acting on it. An agent skill from chrisblattman/claudeblattman.

    463 GitHub stars~1.7k tokensUpdated 2 days ago
    Auto-check passed

Questions about Prompt

What does Prompt do?

Turn a rough request into a clear, structured prompt, show it, then run it. Prompt is an agent skill from chrisblattman/claudeblattman. Turn a rough request into a clear, structured prompt, show it, then run it.

When should I use Prompt?

Prompt fits situations like: your request is rough; one real ask is buried in a long brain dump; you want to refine; improve an existing prompt (refine mode returns the improved prompt without running it).

How do I install Prompt in Claude Code?

Run `npx skills add chrisblattman/claudeblattman --skill prompt -a claude-code`. Or copy the skill folder (plugins/starter-kit/skills/prompt in chrisblattman/claudeblattman) into .claude/skills/prompt in your project. Claude Code loads it when a task matches its description.

How do I install Prompt in Codex?

Run `npx skills add chrisblattman/claudeblattman --skill prompt -a codex`. Or copy the skill folder (plugins/starter-kit/skills/prompt in chrisblattman/claudeblattman) into .agents/skills/prompt in your project. Codex loads it when a task matches its description.

Can I use Prompt in Cursor, Gemini CLI or GitHub Copilot?

Cursor, Gemini CLI, GitHub Copilot and OpenCode also load SKILL.md folders. With the skills CLI, run `npx skills add chrisblattman/claudeblattman --skill prompt -a cursor` (or -a gemini-cli, github-copilot or opencode for the others). To copy it by hand, put the folder in .cursor/skills/prompt, .gemini/skills/prompt, .github/skills/prompt and .opencode/skills/prompt in your project.

What does Prompt need to run?

SKILL.md names no scripts, command-line tools or credentials: Prompt is instructions for the agent only.

Does Prompt access the network?

SKILL.md contains no URLs. Any network use would come from the scripts or tools the agent runs. This is read from the text; nothing was executed.

Is Prompt safe to install?

Our automated static check of SKILL.md found no risky patterns, such as piping downloads into a shell, reading credential files or hidden Unicode. It is not a guarantee. Review the folder before installing.

What licence does Prompt use?

Prompt is published under the MIT licence (the repository's licence). It allows redistribution, so the full SKILL.md is shown on this page.

How many tokens does Prompt use?

About 1.2k tokens (SKILL.md is roughly 4.7k characters). Agents keep only the skill's name and description in context until a task matches; then they load SKILL.md in full. Its references folder adds about 2.4k tokens, read only when the agent opens those files.

What are the alternatives to Prompt?

Skills that share tags, products or a category with Prompt: Ccb Clear (SeemSeam/claude_codex_bridge, 3.6k stars), Ccb Clear (SeemSeam/claude_codex_bridge, 3.6k stars), Ccb Clear (SeemSeam/claude_codex_bridge, 3.6k stars) and Ccb Clear (SeemSeam/claude_codex_bridge, 3.6k stars). The comparison table on this page puts their stars, adoption, token cost, safety result and licence side by side.

Who maintains Prompt?

chrisblattman (a GitHub user) maintains it in chrisblattman/claudeblattman, which has 463 GitHub stars. The repository holds 10 skills in this directory. The repository was last updated on October 7, 2026.

Source: chrisblattman/claudeblattman on GitHub. Facts on this page come from the repository at the commit we read; the author's words are quoted as theirs.