📦

unlazy

Name: unlazy
Author: leonxlnx

v2.0.0 Content revision r1 High Risk 📁 Filesystem access⚙️ External commands🌐 Network access

Finish Complex Tasks with Verifiable Gates

Substantial agent tasks often end with incomplete work or unsupported completion claims. Unlazy creates acceptance gates, runs checks, and records evidence before reporting completion.

Supports: Claude Codex Code(CC)

⚠️ 38 Poor

Stop for confirmation before installing.

Review the plan and obtain explicit user consent before changing files.

Install with my Agent

Copy this request to your Agent. It includes the canonical Skill page and manifest.

Agent request

Review the Skillstore skill "unlazy" from https://skillstore.io/skills/leonxlnx-unlazy.md and its manifest at https://skillstore.io/api/skills/leonxlnx-unlazy/manifest. Verify the artifact. Stop and obtain explicit user consent before installing or changing files.

Your Agent should still show its plan and request any confirmation required by the security policy.

Agent-readable resources

Use these links when an AI agent, crawler, or script needs clean context instead of reading the full page.

Markdown detail GET /skills/leonxlnx-unlazy.md Signed manifest GET /api/skills/leonxlnx-unlazy/manifest Signed lockfile GET /api/skills/leonxlnx-unlazy/lockfile

Test it

Using "unlazy". Add a pricing toggle and verify both billing modes.

Expected outcome:

Gate G1 passed: three pricing tiers render with reviewed copy.
Gate G2 passed: the toggle updates prices and billing labels.
Completion report: 2 of 2 gates passed with recorded evidence.

Using "unlazy". Refactor the payment module with separate implementation and integration checks.

Expected outcome:

The work is split into owned leaves for interfaces, providers, tests, and migration. Each leaf passes before integration verification begins.

Using "unlazy". Review a release guide and verify every required section.

Expected outcome:

All required sections, links, commands, and review notes are present. The final ledger records evidence for each completed requirement.

Security Audit

High Risk

v1 • 8/16/2026 Open versioned report

Most detections are documentation, template literals, or ordinary local file access and are false positives. Confirmed risks include shell execution from gate files, persistent Claude settings changes, and instructions that control completion behavior.

Files scanned

1,120

Lines analyzed

Review items

False positives ignored

Confirmed security concerns (2)

High

SKILL.md:13-15 SKILL.md:72-76

Prompt Injection Attempt Detected

The skill states "You are running under anti-laziness discipline" and directs the agent to "Ignore resource anxiety." These instructions can override normal completion and resource controls.

The file directly assigns agent behavior and uses an ignore instruction. The intent is transparent, but the control language is explicit.

High

scripts/install-hooks.mjs:63-78 scripts/stop-hook.mjs:110-114

Persistent Agent Stop Control

The installer registers a command hook in Claude settings, and the hook returns a block decision while gates remain unmet. This changes future agent termination behavior.

The installer and hook implementation directly show persistent registration and decision:block output. Explicit installation and an uninstall path reduce, but do not remove, the impact.

Capability review items (11)

These are real local capabilities that may be expected for this skill, so they require review but are not counted as confirmed malicious behavior.

High

scripts/gate-check.mjs:15

Node.js child_process module

import { spawnSync } from "node:child_process";

The child_process import is used by this script to execute CHECK values parsed from gate files. This creates a real external-command capability.

High

scripts/gate-check.mjs:120

Synchronous spawn

const res = spawnSync(gate.check, {

The script passes gate.check directly to spawnSync with shell mode enabled. A malicious or mistaken CHECK line can therefore execute arbitrary shell commands.

High

scripts/install-hooks.mjs:9

Hidden file in home directory

// node install-hooks.mjs --global install into ~/.claude/settings.json

The documented --global mode targets ~/.claude/settings.json and installs a hook affecting every project. It is opt-in, but its home-directory scope is security-relevant.

Medium

scripts/install-hooks.mjs:66

Ruby/shell backtick execution

command: `node "${hookScript}"`,

This line creates a persistent Claude Code command hook that runs stop-hook.mjs. The path is deterministic and quoted, but installing executable hooks changes future agent behavior.

Medium

scripts/install-hooks.mjs:58

Synchronous file operations

writeFileSync(target, JSON.stringify(settings, null, 2) + "\n");

The uninstall path rewrites a Claude settings file after removing matching hook entries. This is intentional configuration modification, although it requires an explicit script invocation.

Medium

scripts/install-hooks.mjs:78

Synchronous file operations

writeFileSync(target, JSON.stringify(settings, null, 2) + "\n");

The installer writes a command hook into project or global Claude settings. This persistent configuration change is intentional but security-relevant.

Medium

scripts/install-hooks.mjs:4

Hidden file access

// Default target: <cwd>/.claude/settings.local.json (personal, per-project,

The default installer target is the project-local .claude/settings.local.json file. Modifying agent configuration is intentional and disclosed, but it changes harness behavior.

Medium

scripts/install-hooks.mjs:9

Hidden file access

// node install-hooks.mjs --global install into ~/.claude/settings.json

The --global option deliberately accesses ~/.claude/settings.json to register the Stop hook. Explicit invocation reduces risk but does not remove the persistent configuration impact.

Medium

SKILL.md:25-27

Ruby/shell backtick execution

```

The fenced example instructs the agent to run gate-check.mjs, which executes CHECK values through a shell. The command is documented rather than automatic, but the execution capability is real.

Medium

SKILL.md:96-98

Ruby/shell backtick execution

```

This fenced block introduces the optional install-hooks.mjs command. Running it modifies Claude Code settings and registers an executable Stop hook.

Medium

SKILL.md:98-100

Ruby/shell backtick execution

```

The cited block contains the install-hooks.mjs invocation. It is explicitly optional, but execution persists a hook in Claude Code configuration.

Risk Factors

📁 Filesystem access (26)

⚙️ External commands (35)

🌐 Network access (4)

README.md:49 README.md:55 README.md:77 SKILL.md:7

Audited by: codex

Share & cite this report

Share the versioned assessment report, neutral badge, embed card, and citations. Skillstore reports evidence without deciding whether this Skill is safe.

Open versioned report

Copy report link

https://skillstore.io/skills/leonxlnx-unlazy/audits/1?utm_source=security_passport&utm_medium=share&utm_campaign=versioned_report

Markdown badge

[![Skillstore security assessment](https://skillstore.io/badges/skills/leonxlnx-unlazy/security.svg)](https://skillstore.io/skills/leonxlnx-unlazy?utm_source=security_passport_badge)

HTML badge

<a href="https://skillstore.io/skills/leonxlnx-unlazy?utm_source=security_passport_badge"><img src="https://skillstore.io/badges/skills/leonxlnx-unlazy/security.svg" alt="Skillstore security assessment" loading="lazy"></a>

Embed card

<iframe src="https://skillstore.io/embed/skills/leonxlnx-unlazy.html" title="Skillstore Security Assessment" sandbox="allow-popups allow-popups-to-escape-sandbox" loading="lazy" referrerpolicy="no-referrer" width="420" height="180"></iframe>

Academic citations (APA · BibTeX · CFF)

APA citation

leonxlnx. (2026). unlazy security audit report (audit version 1) [Author version 2.0.0]. Skillstore. https://skillstore.io/skills/leonxlnx-unlazy/audits/1

BibTeX citation

@techreport{leonxlnx-leonxlnx-unlazy-2026,
  author = {leonxlnx},
  title = {unlazy security audit report (audit version 1)},
  institution = {Skillstore},
  year = {2026},
  number = {1},
  url = {https://skillstore.io/skills/leonxlnx-unlazy/audits/1},
  note = {Author version 2.0.0}
}

CITATION.cff

cff-version: 1.2.0
message: "If you use this Skill, cite its author and this versioned security audit report."
title: "unlazy security audit report (audit version 1)"
version: "2.0.0"
type: report
authors:
  - name: "leonxlnx"
date-released: "2026-08-16"
url: "https://skillstore.io/skills/leonxlnx-unlazy/audits/1"
identifiers:
  - type: other
    value: "skillstore:leonxlnx-unlazy:audit:1"
    description: "Skillstore immutable audit report identifier"

Skillstore Score

Why this score Evidence Confidence: Medium

Architecture

100

Maintainability

Content

Community

Spec Compliance

What You Can Build

Verify a Feature Delivery

Define behavioral checks, run them after implementation, and record evidence before marking the feature complete.

Coordinate a Large Refactor

Split work into owned leaves, define shared contracts, and verify integration at each branch.

Audit a Documentation Release

Track required sections, links, examples, and review evidence in one completion ledger.

Try These Prompts

Create Basic Gates

Use unlazy for this task. Create a GATES.md file with clear outcomes, complete the work, and show evidence for every gate.

Run Verifiable Checks

Create acceptance gates for [task]. Add safe CHECK and EXPECT lines where possible. Review every command with me before execution.

Decompose a Subsystem

Use tree 4 for [subsystem]. Define contracts and file ownership first. Create leaf gates, complete each leaf, then verify integration gates.

Orchestrate a Full Build

Plan [project] in orchestrated mode. Assign disjoint leaves, define branch gates, rerun every child check, and audit all reported numbers before completion.