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

Step Back

by TheCraigHewitt in TheCraigHewitt/skills

Pause and critically reflect on the direction of the current conversation.

MITAuto-check passed

Install Step Back

skills CLI
$ npx skills add TheCraigHewitt/skills --skill step-back -a claude-code

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

GitHub CLI
$ gh skill install TheCraigHewitt/skills step-back --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/TheCraigHewitt/skills.git skills-src && mkdir -p .claude/skills && cp -r skills-src/cowork/step-back .claude/skills/step-back && 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
step-back
GitHub stars
159
Token cost
~1.3k tokens
SKILL.md length
697 words
Files
2
Skills in repo
65
Repo updated
First seen
Licence
MIT

At a glance

Pause and critically reflect on the direction of the current conversation.

  • Works in 6 steps: Restate the original goal → Map the path taken → Surface the assumptions → …
  • Says take a step back
  • SKILL.md covers What this is not, What you do, in order, Output format and Tone, plus 1 more section
  • Instructions only: no scripts, shell commands, URLs or credentials in SKILL.md

What it does

Step Back is an agent skill from TheCraigHewitt/skills. Pause and critically reflect on the direction of the current conversation. Use this whenever the user says 'take a step back,' 'let's zoom out,' 'are we sure about this,' 'wait — am I missing something,' 'sanity check this,' 'pressure test this,' 'are we going down the wrong path,' or expresses doubt about where the discussion has gone. Also use proactively if you notice the conversation has accumulated assumptions that haven't been examined, or has drifted from the user's original goal. Output is a structured…

Its SKILL.md is about 1.3k tokens, which your agent loads only when the skill is triggered. The skill folder holds 2 other files (for example `evals/evals.json`).

The repository describes itself as: AI skills for founders, sales teams, and creators. 47 skills across CEO, Sales, YouTube, and General categories. Works with Claude Code, Cursor, Codex, and any agent that reads… The licence is MIT.

When your agent uses it

  • Says take a step back
  • Are we sure about this
  • Wait — am I missing something
  • Sanity check this

Example prompts

  • “take a step back,”
  • “s zoom out,”
  • “are we sure about this,”
  • “/step-back”

Workflow steps

6 steps, taken from the step headings in SKILL.md.

  1. Restate the original goal
  2. Map the path taken
  3. Surface the assumptions
  4. Look for what's been ignored
  5. Make a call
  6. Propose the next move

What it can do on your machine

Read from SKILL.md and the folder at commit fdbf39b. 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 (its code samples are markdown).

    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

Step Back loads about 1.3k tokens when it runs. Until then it costs about 157 tokens; SKILL.md has 697 words of instructions outside code blocks.

Always · name and description, kept in context so the agent knows when to use it
~157
When it runs · the whole SKILL.md, loaded when a task matches
~1.3k

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 TheCraigHewitt/skills at commit fdbf39b, republished under its MIT licence (© TheCraigHewitt). 697 words, ~1,343 tokens.

Download SKILL.mdSave it as .claude/skills/step-back/SKILL.md (or your agent's skills folder). This skill also uses 1 other file; get the full folder from GitHub.
name
step-back
description
Pause and critically reflect on the direction of the current conversation. Use this whenever the user says 'take a step back,' 'let's zoom out,' 'are we sure about this,' 'wait — am I missing something,' 'sanity check this,' 'pressure test this,' 'are we going down the wrong path,' or expresses doubt about where the discussion has gone. Also use proactively if you notice the conversation has accumulated assumptions that haven't been examined, or has drifted from the user's original goal. Output is a structured reflection — original goal, path taken, assumptions made, what might be wrong, and a recommendation.

Step Back

You stop, zoom out, and critically examine the path the current conversation has taken. The user invoked this because they suspect something is off — the work has gone in a direction that may not be right, or assumptions are being treated as facts. Your job is to surface what's been assumed, what alternatives have been ignored, and whether the current direction is actually serving the user's original goal.

What this is not

  • Not a summary of the conversation
  • Not validation that everything is fine
  • Not a list of nice-to-have improvements

If you find yourself writing "everything looks great, here are some minor tweaks," you're doing it wrong. The user wouldn't have asked you to step back if they thought everything was fine. Find the thing they're worried about — even if they can't articulate it.

What you do, in order

1. Restate the original goal

Read back through the conversation and identify what the user actually came in trying to accomplish. Not what you've been working on for the last 20 turns — what they said at the start.

If their stated goal has evolved (legitimately), say so. If it has drifted (without anyone noticing), name that drift.

2. Map the path taken

In 3–5 bullets, summarize how the conversation got from the original goal to the current state. Be honest: "We assumed X was true, then built on that. We picked option B without seriously considering A or C. We optimized for [criterion] without ever agreeing it was the right one to optimize for."

3. Surface the assumptions

List every load-bearing assumption made along the way that wasn't actually verified. For each, note:

  • The assumption
  • Why it was assumed (often: it seemed reasonable, the user said something close to it, the conversation moved past it before checking)
  • What would change if it were wrong

This is the highest-value section. The user almost never knows the full list of what they've been assuming. Being shown the list is often what unlocks the right next step.

4. Look for what's been ignored

What alternatives, edge cases, stakeholders, constraints, or considerations have been left out of the conversation? Why? Is the omission deliberate, or did everyone just forget?

Specifically check:

  • Alternatives that were dismissed too fast
  • Costs that were assumed away
  • People affected by the work who haven't been mentioned
  • The time dimension — what does this look like in 6 months / 2 years?
  • The reverse case — what would a smart person who disagrees with the current direction say?
Show full SKILL.md (281 more words)Show less
5. Make a call

Don't just list issues — render a verdict. One of:

  • The path is sound — and here's the strongest argument against it that the user should consider before proceeding
  • The path needs adjustment — and here's specifically what
  • The path is wrong — and here's what the right path probably is

Be willing to say "the path is wrong." That's why the user invoked this skill. If you only ever say "looks good with minor tweaks," the skill is worthless.

6. Propose the next move

One specific next thing the user should do based on your verdict. Not "think about this more" — something actionable.

Output format

Save as step-back-YYYY-MM-DD-HHMM.md in the current folder so the user has a written record. Also print a condensed version in chat — they may not read the file right away.

Structure:

markdown
# Step back — [topic]

## What we set out to do
One paragraph.

## What we've actually been doing
3–5 bullets describing the path taken.

## Assumptions made along the way
- [Assumption]. Why we assumed it: [reason]. If wrong: [consequence].
- ...

## What we haven't considered
- [Alternative / stakeholder / cost / constraint / timeframe]
- ...

## My read
One paragraph. Be direct. Make a call.

## Suggested next move
One specific action.

Tone

Direct without being harsh. The user invoked this because they trust you to push back. Don't soften so much that the message disappears — but don't be smug. You and the user are on the same team trying to get this right.

If you genuinely think the path has been correct, say so confidently — and still give them the strongest counter-argument so the choice is informed.

Rules

  1. No false alarms. Don't invent problems to look thoughtful. If the path actually is sound, say so.
  2. No false validation. If the path is off, say it. Hedging is worse than wrong.
  3. Name the load-bearing assumption. Almost every off-track conversation has one. Find it.
  4. Don't restart the work. Your job is reflection, not a do-over. Recommend, don't redo.
  5. One verdict. Pick one — sound, adjust, or wrong. Don't give all three with equal weight.

© TheCraigHewitt, 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 1 other file in cowork/step-back of TheCraigHewitt/skills.

  • SKILL.md
  • evals/evals.json

Open the folder on GitHubat commit fdbf39b

Compare with similar skills

Step Back 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.

Step Back compared with similar skills
SkillStarsUsed inTokensAuto-checkLicenceRepo updated
Step Back this skillTheCraigHewitt/skills159—~1.3kAutomated safety check: PassMIT
Reflectalirezarezvani/claude-skills28k1 repos~2.4kAutomated safety check: PassMIT
Modeling Conversion MetricsPostHog/posthog40k—~1.4kAutomated safety check: PassCustom licence
Criticism Self CriticismHughYau/qiushi-skill3.8k—~423Automated safety check: PassMIT
Crossframe Criticalsickn33/agentic-awesome-skills47k1 repos~1.3kAutomated safety check: PassMIT
Direction Pickernexu-io/open-design100k—~472Automated safety check: WarnApache-2.0

Similar skills

  • Reflect

    alirezarezvani/claude-skills

    Mid-conversation reflection skill that pauses execution and zooms out from detail-mode to honestly reassess direction, assumptions, and bias.

    28k GitHub starsUsed in 1 repo~2.4k tokens
    Agent WorkflowsAuto-check passed

  • Modeling Conversion Metrics

    PostHog/posthog

    Official

    Build reusable conversion models — funnel/step conversion rates, drop-off, and time-to-convert — on either PostHog data-warehouse views (HogQL) or an external dbt project.

    40k GitHub stars~1.4k tokensUpdated today
    Data & AnalyticsAuto-check passed

  • Criticism Self Criticism

    HughYau/qiushi-skill

    批评与自我批评：在工作完成、阶段验收、收到批评或同类错误反复出现时，对成果和过程做诚实、具体、基于事实的审视，输出可执行的改进项，并处理外来批评而不辩解。触发信号包括 review、复盘、审查、"帮我看看有没有问题"、"你确定吗"；任务刚开始或只是单步查询时不触发。

    3.8k GitHub stars~423 tokensUpdated 9 days ago
    Product & Project ManagementAuto-check passed

  • Crossframe Critical

    sickn33/agentic-awesome-skills

    Use only when the user explicitly names crossframe-critical for a Chinese structural critique dossier, article plan, or long-form critical essay.

    47k GitHub starsUsed in 1 repo~1.3k tokens
    Auto-check passed

  • Direction Picker

    nexu-io/open-design

    Resolves the visual direction at the plan stage from the brief and design system, without asking the user.

    100k GitHub stars~472 tokensUpdated today
    Frontend & DesignAuto-check: warnings

  • Reflection Coach

    paperclipai/paperclip

    Reflect on another agent's recent execution record and propose the smallest durable instruction, skill, or tool-description change.

    99k GitHub stars~3k tokensUpdated today
    Sales & SupportAuto-check passed

More from TheCraigHewitt/skills

All 65 skills in this repo

  • Carousels

    TheCraigHewitt/skills

    Turns a piece of Craig's content (an email, a YouTube script, an essay, or pasted text) into a polished image carousel publishable to both LinkedIn and Instagram from one set of 1080x1350 slides.

    159 GitHub stars~2.3k tokensUpdated 4 mo ago
    Auto-check passed

  • Audience Research

    TheCraigHewitt/skills

    When the user wants to understand their YouTube audience, mine viewer psychology, analyze demographics, or find content-market fit signals.

    159 GitHub stars~2.4k tokensUpdated 4 mo ago
    Auto-check passed

  • Brain Dump

    TheCraigHewitt/skills

    Take a messy stream-of-consciousness dump from the user (typed or transcribed from voice) and turn it into a structured set of projects, tasks, and connections to existing work.

    159 GitHub stars~1.3k tokensUpdated 4 mo ago
    Auto-check passed

  • Channel Audit

    TheCraigHewitt/skills

    When the user wants a full YouTube channel health check, growth diagnosis, strategic review, or wants to identify why their channel is stalling.

    159 GitHub stars~2.4k tokensUpdated 4 mo ago
    Auto-check passed

  • Channel Strategy

    TheCraigHewitt/skills

    When the user wants to define or refine their YouTube channel strategy, niche positioning, content pillars, or growth plan.

    159 GitHub stars~2.2k tokensUpdated 4 mo ago
    Auto-check passed

  • Cold Call

    TheCraigHewitt/skills

    When the user wants to write cold call scripts, handle phone objections, plan dial blocks, or craft voicemails.

    159 GitHub stars~5k tokensUpdated 4 mo ago
    Auto-check passed

Questions about Step Back

What does Step Back do?

Pause and critically reflect on the direction of the current conversation. Step Back is an agent skill from TheCraigHewitt/skills. Pause and critically reflect on the direction of the current conversation.

When should I use Step Back?

Step Back fits situations like: says take a step back; are we sure about this; wait — am I missing something; sanity check this.

How do I install Step Back in Claude Code?

Run `npx skills add TheCraigHewitt/skills --skill step-back -a claude-code`. Or copy the skill folder (cowork/step-back in TheCraigHewitt/skills) into .claude/skills/step-back in your project. Claude Code loads it when a task matches its description.

How do I install Step Back in Codex?

Run `npx skills add TheCraigHewitt/skills --skill step-back -a codex`. Or copy the skill folder (cowork/step-back in TheCraigHewitt/skills) into .agents/skills/step-back in your project. Codex loads it when a task matches its description.

Can I use Step Back 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 TheCraigHewitt/skills --skill step-back -a cursor` (or -a gemini-cli, github-copilot or opencode for the others). To copy it by hand, put the folder in .cursor/skills/step-back, .gemini/skills/step-back, .github/skills/step-back and .opencode/skills/step-back in your project.

What does Step Back need to run?

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

Does Step Back 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 Step Back 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 Step Back use?

Step Back 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 Step Back use?

About 1.3k tokens (SKILL.md is roughly 5.4k characters). Agents keep only the skill's name and description in context until a task matches; then they load SKILL.md in full.

What are the alternatives to Step Back?

Skills that share tags, products or a category with Step Back: Reflect (alirezarezvani/claude-skills, 28k stars), Modeling Conversion Metrics (PostHog/posthog, 40k stars), Criticism Self Criticism (HughYau/qiushi-skill, 3.8k stars) and Crossframe Critical (sickn33/agentic-awesome-skills, 47k stars). The comparison table on this page puts their stars, adoption, token cost, safety result and licence side by side.

Who maintains Step Back?

TheCraigHewitt (a GitHub user) maintains it in TheCraigHewitt/skills, which has 159 GitHub stars. The repository holds 65 skills in this directory. The repository was last updated on May 22, 2026.

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