przeprogramowani/10x-cli/10x-cli-guide

Claude Code and agent skill by przeprogramowani/10x-cli

Trust tier 2
  • community
LLM review pending
Source check pending

Checking that the SKILL.md file is still in the repository.

Use when the user wants to download, use or update 10xDevs CLI skills, choose a helper installation channel, inspect course content, switch tool profiles, see what changed in 10x-toolkit releases (changelog) or troubleshoot CLI/auth/content conflicts. Guides filtered get → an actual agent task → sync while preserving local work and course/tool/language context. For first installation or authentication preparation, use an available 10x-cli-setup copy. Does not implement CLI runtime or grant cour…

Install

npx vskill@latest install przeprogramowani/10x-cli/10x-cli-guide

SKILL.md

10x-cli: download, use, update

Read the bundled compatibility and channel reference before issuing commands. It contains the pinned runner setup, version checks, both helper channels and ownership safeguards. Commands use the released lesson-scoped skill filter; verify the actual selected package and content before using it. A local build or source membership is not proof that a feature has shipped.

The next CLI release also provides project-only bundled installation through 10x helpers install --tool <chosen-profile>. This command is unreleased and absent from the 1.21.0/1.22.0 master baselines: check the actual runner's helpers --help first. Follow Bundled public copies in the local reference for complete files, explicit targets and conflict handling; keep the existing pinned public route when the runner does not support it.

Environment

Reuse the setup handoff: project root, course, tool, language, runner/version, auth/access status, update method and helper channel/path. If anything is absent, inspect only that item. A working installed CLI needs no reinstall. If setup is needed, locate its actual SKILL.md and references or install that helper through the public channel; do not invoke a missing sibling by name.

The guided acceptance context is macOS/zsh, Claude Code, 10xdevs4 and Polish. Determine the actual OS and shell from the environment, not a POSIX command that labels every failure Windows. Select the intended project root before any write. For v4, retain an existing v3 project and use a separate directory; ordinary get, sync and profile changes do not migrate editions. Preserve .10x-cli.json and all manifests if their versions/courses conflict.

Use the verified 10x_cli runner defined in the reference, or the user's exact verified global/standalone executable:

bash
10x_cli --version
10x_cli get --help
10x_cli sync --help
10x_cli auth --status
10x_cli list

Leave --course out of these commands: the CLI reads the edition from .10x-cli.json when the directory is already bound, otherwise from the live API recommendation (the highest edition the account can reach). get prints Course: <slug> (<reason>); check it once before the first write. Use --course <slug> only to inspect another entitled edition on purpose — a bound project cannot be rewritten to a different edition (course_mismatch), so v4 starts in a new empty directory.

Check source/release evidence for lesson reference, skill filter and lesson-scoped sync, then use the filtered preview below to check the endpoint. Do not treat a successful help exit as capability proof. Keep unsupported CLI syntax, unpublished/missing content, locked module, membership denial and network failure distinct. Missing final release evidence need not block preparing the public helpers or the exercise files.

Read only needed nonsecret preferences from config.json. On macOS/Linux its base is nonempty $XDG_CONFIG_HOME, otherwise ~/.config; on Windows it is %APPDATA%, otherwise the user's AppData/Roaming. Append 10x-cli/config.json. Do not print auth.json, discard stderr, truncate doctor JSON or erase config to repair an unknown problem. An explicit course/tool/language in this journey takes precedence over saved defaults for that command.

Session management

When login is needed, let the user choose the delivery channel and complete it:

bash
10x_cli auth # Interactive: choose email or Circle
10x_cli auth --method email # Email magic link; default for piped/JSON output
10x_cli auth --method circle # One-time approval link delivered in Circle
10x_cli auth --status
10x_cli auth --logout

Circle is useful when the email does not arrive. The approval link expires after 15 minutes; the CLI does not resend it automatically. In non-interactive mode, provide the user's email with --email and choose --method circle explicitly. Do not send a login message until the user requests authentication. Sessions refresh transparently; re-login is needed only when refresh cannot recover them.

Download

The launch exercise follows lesson 1, “Od pomysłu do PRD”, using its existing 10xCards example. 10x-plan is not part of this launch demonstration. Prepare all four lesson skills: 10x-idea-check, 10x-init, 10x-shape and 10x-prd. Idea assessment is optional to run; its files should still be available when preparing lesson 1. If the user requested only a specific skill, retain that narrower scope. After checking each name's capability and availability, download four separate complete selected skill trees. Inspect each dry-run before its corresponding write:

bash
10x_cli get m1l1 --type skills --name 10x-idea-check --tool claude-code --lang pl --dry-run
10x_cli get m1l1 --type skills --name 10x-idea-check --tool claude-code --lang pl
10x_cli get m1l1 --type skills --name 10x-init --tool claude-code --lang pl --dry-run
10x_cli get m1l1 --type skills --name 10x-init --tool claude-code --lang pl
10x_cli get m1l1 --type skills --name 10x-shape --tool claude-code --lang pl --dry-run
10x_cli get m1l1 --type skills --name 10x-shape --tool claude-code --lang pl
10x_cli get m1l1 --type skills --name 10x-prd --tool claude-code --lang pl --dry-run
10x_cli get m1l1 --type skills --name 10x-prd --tool claude-code --lang pl

Inspect each report and the complete supporting tree, not just SKILL.md. In the chosen macOS/zsh exercise directory, each of these checks must succeed before use (stop on any failure; do not infer success from the last check alone):

bash
test -s .claude/skills/10x-idea-check/SKILL.md
test -s .claude/skills/10x-idea-check/references/examples.md
test -s .claude/skills/10x-idea-check/references/assessment-guide.md
test -s .claude/skills/10x-idea-check/references/10xdevs-4-dates.md
test -s .claude/skills/10x-idea-check/references/10xdevs-4-certification.md
test -s .claude/skills/10x-init/SKILL.md
test -s .claude/skills/10x-shape/SKILL.md
test -s .claude/skills/10x-shape/references/prd-schema.md
test -s .claude/skills/10x-prd/SKILL.md
test -s .claude/skills/10x-prd/../10x-shape/references/prd-schema.md

The PRD entrypoint resolves ../10x-shape/references/prd-schema.md relative to its own directory. A standalone PRD tree is insufficient. Read all installed entrypoints and every reference they require; the paths above are the known source minimum, not permission to discard extra files from a published bundle. Also inspect .claude/.10x-cli-manifest.json: lessons.m1l1.skills must include all four names, with file hashes in files.skills. These are lesson-owned partial downloads, not independent owners. Inspect .10x-cli.json for the course binding; partial downloads do not establish a complete lesson release identity.

CLI 1.21.0 is published with v4 and filtered skill downloads; production m1l1 EN/PL contains idea-check/init/shape/prd and their references. These revised helpers are a separate source change, not proof that their course copies have been published. Verify all four names against the actual selected release. If any name, schema, owner or release is missing/mismatched, preserve the precise error and stop the exercise; never silently substitute a whole lesson, another course or filtered get.

CLAUDE-m1l1 is a separate lesson rule, not delivered by these filtered skill gets. The inspected init/shape/prd sources do not require that rule to run this chain. This does not establish that every step of the full lesson works without it. Use the learner's existing lesson 1 inputs and instructions: if they require the rule, inspect the existing project rule and its provenance. If absent, report the missing prerequisite and obtain the supported route from the lesson/release owner before that step. Do not invent a rule command or download a full lesson to bypass it. Do not overwrite an existing project rule.

For browsing use 10x_cli list m1. The commands above filter one lesson by skill name; get 10x-init is not supported. --print is inspection: TTY Markdown can contain only SKILL.md; non-TTY output is a JSON envelope. Never redirect print output into SKILL.md as a package installation.

Use: 10xCards, from idea to PRD

Downloading the trees is only preparation. Use the existing 10xCards example and the learner's actual answers from lesson 1. If those inputs are absent, ask for them; do not invent product requirements, a replacement task.md or a ready-made plan. Keep private lesson text out of public fixtures and transcripts. Do not assume native slash/$ discovery or automatic activation from npm install. Give the agent explicit local paths. If the learner wants to assess whether an idea fits their experience, time and course goals, first read .claude/skills/10x-idea-check/SKILL.md and its references and follow that skill. Do not make assessment a prerequisite when the learner is ready to shape. Then work through these steps separately:

  1. Read .claude/skills/10x-init/SKILL.md and follow it in the chosen project.

Inspect the create-if-absent context/changes, context/archive and context/foundation directories and their READMEs. Preserve existing files.

  1. Read .claude/skills/10x-shape/SKILL.md and

.claude/skills/10x-shape/references/prd-schema.md. Follow the skill's discovery with the learner's 10xCards inputs. Let the learner answer and approve the checkpoint; do not answer for them. Inspect context/foundation/shape-notes.md against those answers before proceeding.

  1. Read .claude/skills/10x-prd/SKILL.md and its sibling schema, then generate the

draft from the actual context/foundation/shape-notes.md. Inspect context/foundation/prd.md against that input and the installed schema; unresolved domain choices stay open. Respect the skill's existing-file collision choice (a versioned file may be the appropriate result).

Success requires the learner's notes and a schema-conformant PRD, with gaps explicit and original project work preserved. A transcript of downloads alone is insufficient. Stop after reviewing the PRD; do not chain into stack selection, bootstrap or implementation. If a different global/local skill copy was read, correct the path before accepting the result. Record the actual agent, profile, language, output paths and checks; use a fresh isolated exercise directory for the two canonical output filenames instead of forcing an overwrite.

Update

Use the same runner, directory, tool and language; the edition stays the project's own, so sync needs no --course either. Sync updates entire downloaded lessons, including m1l1 after these filtered gets. Its preview may include other skills, prompts, configs and course rules. Inspect that expanded scope and apply only when the user accepts it; to update only one skill, repeat its filtered preview/get instead. Do not use sync as a hidden rule prerequisite workaround:

bash
10x_cli sync --tool claude-code --lang pl --dry-run
10x_cli sync --tool claude-code --lang pl

Normal sync refreshes the full lessons recorded in the manifest, not just the four selected skills. --all broadens scope to unlocked lessons and is not needed for this exercise. Missing managed files should be repaired; local edits should remain visible as conflicts or preserved files. Read all report outcomes and resource counts even if exit is 0: skipped conflicts alone are not process errors. Do not equate an unchanged remote digest with intact local files.

This is the start of the file. Read the full SKILL.md on GitHub.

Security

  • Security scanPassed, 100 of 100Pattern checks on the skill's files.
  • Model reviewNot runA model review is scheduled for the next rescan.

Quality evaluation

  • Task evaluationNot runNo evaluation yet. Run one in Skill Studio.

Source

Popularity

Repository stars
79
Repository forks
19
Repository npm downloads, 7 days
1,989
7-day trend score
32

Stars and forks are for the whole GitHub repository, not this skill alone. Downloads are for the repository's npm package. Refreshed 20 May 2026, 5 months ago.

Works with

Any agent vskill installs to: 56 targets in all. These 39 have a page on this site.

  • Claude Code
  • Cursor
  • GitHub Copilot
  • Windsurf
  • Codex
  • Gemini CLI
  • Cline
  • Amp
  • Roo Code
  • Goose
  • Continue
  • Kilo Code
  • Junie
  • Trae
  • OpenHands
  • Qwen Code
  • AI Agent
  • Tabnine Chat
  • Sourcegraph Cody
  • Amazon Q
  • Firebase Genkit
  • Sweep
  • Devin
  • Phind
  • Mentat
  • Claude Desktop
  • Aider
  • GPT Pilot
  • Plandex
  • bolt.new
  • v0
  • Replit Agent
  • Anse
  • TypingMind
  • BetterChatGPT
  • LibreChat
  • Warp AI
  • Pieces
  • Fig AI
Embed badge
vskill VERIFIED badge
[![vskill VERIFIED](https://verified-skill.com/api/v1/skills/przeprogramowani/10x-cli/10x-cli-guide/badge)](https://verified-skill.com/skills/przeprogramowani/10x-cli/10x-cli-guide)

Something wrong with this skill or its page? Tell us.