howdeploy

Pizdec — Security skill for Claude Code

Security community

Read-only, evidence-driven security audit skill for AI coding agents.

How to install Pizdec

This entry records only its repository, not the path inside it, so there is no exact command to give. Open howdeploy/pizdec and copy the folder into ~/.claude/skills/, or the file into ~/.claude/agents/.

What Pizdec does

Read-only, evidence-driven security audit skill for AI coding agents.

Alternatives in Security

Deepsec — Deepsec is a security harness for finding vulnerabilities in your codebase powered by coding agents 7.8k ★
Claude Bootstrap — Opinionated project initialization with security-first guardrails, spec-driven atomic todos, LLM testing patte 536 ★
Maestro Knowledge — Intent-driven knowledge-store and Run knowledge lifecycle management — audit/prune, stage candidates (with sig 530 ★

README

PIZDEC — Project & Infrastructure Zero-trust Defense Exposure Check

*Find it before it finds you.*

`PIZDEC` is both the acronym above and a deliberate transliteration of the Russian word «пиздец»: what an overlooked security exposure can turn into.

[Русская версия](README.ru.md)

PIZDEC is a platform-neutral, read-only security-audit skill for AI coding agents. It instructs an agent to connect code to its deployed environment, inspect the selected depth systematically, and produce a short evidence-based report for both a human and a future agent.

PIZDEC does not fix anything. The distributable `pizdec/` directory contains no installer, executable audit scripts, telemetry, remote fetches, universal shell commands, or prewritten remediation. The repository's separate `evals/` directory includes only a read-only fixture validator.

What it audits

Websites, frontends, backends, APIs, webhooks, and bots.
Servers, workstations, identities, SSH, services, listeners, firewalls, tunnels, and persistence.
Docker, Compose, Kubernetes, images, mounts, volumes, networks, and registries.
Databases, caches, queues, object storage, backups, and direct data-access paths.
First-party source code, configuration, infrastructure, migrations, hooks, and deployment trees.
CI/CD, dependencies, package lifecycle, Git exposure, and software supply chain.
AI agents, prompts, memory, skills, plugins, MCP tools, browser control, and excessive agency.

Every mode starts with breadth discovery and checks dangerous trust paths early. The selected mode controls how far code and lower-risk surfaces are reviewed after that.

Modes

User-facing mode	Machine mode	Intended use	Coverage result	Changes the target?
PIZDEC Triage	`TRIAGE`	Fast check of exposed infrastructure, identities, ports, containers, agents, bots, and one or two risk-ranked projects	`TRIAGE_COMPLETE` or `PARTIAL_READ_ONLY`	No
PIZDEC Full	`FULL`	Comprehensiv