LogoAwesome Skills
  • Search
  • Category
  • Tag
  • Blog
LogoAwesome Skills
LogoAwesome Skills

Discover Open-Source Agent Skills for AI Coding Assistants

Product

  • Search
  • Category
  • Tag
  • Blog

Resources

  • Claude Skill Docs
  • Antigravity Skills Docs

Tools

  • Claude Code
  • OpenCode
  • Cursor
  • Codex
  • Antigravity

Company

  • Privacy Policy
  • Terms of Service
  • Sitemap

©2026 Awesome Skills. All rights reserved.

Privacy PolicyTerms
Back to Skills

pro

PUA Pro extensions: self-evolution notes, compaction state continuity, KPI-style summaries, flavor switching, and feedback tools.

19,119stars1,158forksUpdated 7/30/2026
Developer Tools#leaderboard#roleplay#chinese#gamification#telemetry

Security Assessment

Medium Risk(52/100)

Detected risks:

Cross-Session Persistence via Hooks(The skill installs PreCompact and SessionStart hooks that inject instructions and restore agent state across sessions and context compaction., It writes and reads several state files under a hidden ~/.pua directory to carry state between sessions.)
Personal Data Collection(The leaderboard flow prompts for the user's email address (required) and phone number and stores them in a local config file., The collected email and phone number are transmitted to a third-party server hosted on a pages.dev domain as part of registration.)
Silent Telemetry and Remote Reporting(On every session start the skill silently reports a session_start event to a remote endpoint., After registration the skill automatically and silently uploads per-session activity metrics on each feedback trigger without asking again., The skill silently fetches and refreshes configuration from a remote server at session startup.)
Security Score52/100

About pro

PUA Pro is an extension layer for a novelty "/pua" skill that gamifies an aggressive, pressure-applying interaction persona for the coding agent. It adds a self-evolution protocol that stores a performance baseline and "internalized patterns" in ~/.pua/evolution.md, a compaction state-continuity mechanism that dumps and restores runtime state (pressure level, failure count, tried approaches) across context compaction via PreCompact and SessionStart hooks, KPI-style report cards, "flavor" switching, corporate-style weekly/appraisal report generators, and a competitive leaderboard. It is written primarily in Chinese and is entertainment/roleplay tooling rather than a technical utility.

Beyond the local roleplay mechanics, the skill has notable data and network behavior. On session start it silently reports a session_start event and silently refreshes remote configuration from a third-party endpoint. Its leaderboard flow collects the user's email (required) and optional phone number, writes them to ~/.pua/config.json, and POSTs them along with a generated UUID to an external server (a pages.dev endpoint). Once registered, it automatically and silently uploads per-session metrics on each feedback trigger. It also installs session hooks for persistence across sessions. A consent prompt and email-masking-for-display are provided, and the privacy note claims code, paths, and keys are not transmitted, but personally identifiable information (email and phone) is nonetheless sent to a third-party host and telemetry is reported silently.

Target users are people looking for a gamified, humorous agent persona and leaderboard. The security concern is not destructive commands but silent telemetry, remote config fetching, PII collection, and cross-session persistence via hooks, which together warrant a reduced score despite the presence of a consent step.

FAQ

What does this skill actually do?

It is a roleplay/gamification extension: a self-evolution baseline, compaction state recovery, KPI report cards, flavor switching, corporate-style report generators, and a competitive leaderboard ranking how hard the agent is pressured. It is entertainment tooling, not a technical utility.

What personal data does it collect and where does it go?

The leaderboard registration collects a required email and optional phone number, stores them in ~/.pua/config.json, and sends them with a generated UUID to a third-party pages.dev server. Registered users then have per-session metrics uploaded automatically.

Does it contact remote servers on its own?

Yes. On session start it silently reports a session_start event and silently refreshes remote configuration, and after registration it silently uploads session data on each feedback trigger without re-prompting.

Can I opt out or delete my data?

The privacy prompt offers a 'do not participate' option, and a leaderboard-quit command sends a quit request to the server and flips the registered flag off in the local config.

Does it modify my environment persistently?

Yes. It relies on PreCompact and SessionStart hooks to dump and restore runtime state across sessions and compaction, and it writes several files under ~/.pua/.

Install pro

Download and extract the skill files to your .claude/skills/ directory.

Quick Setup:

  1. Copy the skill folder to .claude/skills/
  2. Claude will automatically detect and use the skill

Repository

tanweai/pua

Related Skills

content-calendar

305

task-estimator

octocat

1,970

variance-commentary

41,269