Agentic Skills

High-agency agent skills for turning rough intent into executable specs and shipped work.

This repository contains two companion skills:

• $spec: deep executable spec builder.
• $ship: extreme implementation mode.

Use $spec when the work is fuzzy and needs to become a concrete build contract. Use $ship when the work is ready to execute and should be driven through implementation, verification, polish, and handoff.

Skills

Skill	Purpose	Use When
spec	Convert rough ideas into implementation-ready specs with task trees, acceptance criteria, and verification plans.	You need a PRD, design doc, technical plan, task breakdown, requirements, or $ship handoff.
ship	Execute end-to-end with high agency: inspect, implement, work through blockers, verify, polish, and report.	You want an AI agent to build, debug, research, verify, and keep moving until the requested outcome is handled.

Install

Windows PowerShell

git clone https://github.com/Rtx09x/agentic-skills.git
cd agentic-skills
.\scripts\install.ps1

macOS / Linux

git clone https://github.com/Rtx09x/agentic-skills.git
cd agentic-skills
./scripts/install.sh

These skills are plain Markdown skill folders with optional agent UI metadata. They are currently packaged in a Codex-compatible layout and can be adapted to any agent runtime that supports reusable prompt/skill files.

By default, the install scripts target Codex-compatible skill directories:

• $env:CODEX_HOME\skills on Windows when CODEX_HOME is set.
• $CODEX_HOME/skills on Unix when CODEX_HOME is set.
• ~/.codex/skills otherwise.

To install somewhere else:

.\scripts\install.ps1 -Destination "C:\path\to\skills"

./scripts/install.sh /path/to/skills

Manual Install / Agent Adaptation

For Codex-compatible agents, copy each skill folder into your skills directory:

skills/spec -> ~/.codex/skills/spec
skills/ship -> ~/.codex/skills/ship

Each skill folder must keep this shape:

skill-name/
+-- SKILL.md
+-- agents/
    +-- openai.yaml

For other agents, copy the relevant SKILL.md content into that agent's reusable instruction, rule, prompt, or skill mechanism. The files are intentionally self-contained and do not require Codex-specific tools at runtime.

Usage

Use $spec to shape the work:

Use $spec to turn this product idea into an implementation-ready spec with a task tree, acceptance criteria, verification plan, and $ship handoff.

Use $ship to execute:

Use $ship to implement this end-to-end, work through blockers, verify it, and report exactly what changed.

Recommended workflow:

1. Run $spec for new, ambiguous, high-stakes, or multi-step work.
2. Review the generated task tree and acceptance criteria.
3. Run $ship to execute the plan.
4. Require final reporting with exact files, commands, verification status, and remaining risk.

Philosophy

These skills are built around a simple operating model:

• own the outcome,
• inspect the real system,
• research current facts when uncertain,
• decompose big work into executable tasks,
• keep moving through blockers,
• verify honestly,
• report with receipts.

$spec optimizes for clarity. $ship optimizes for execution.

Repository Layout

.
+-- skills/
|   +-- spec/
|   |   +-- SKILL.md
|   |   +-- agents/openai.yaml
|   +-- ship/
|       +-- SKILL.md
|       +-- agents/openai.yaml
+-- scripts/
|   +-- install.ps1
|   +-- install.sh
|   +-- validate-layout.ps1
+-- docs/
|   +-- usage.md
+-- releases/
|   +-- v0.1.0.md
+-- CHANGELOG.md
+-- LICENSE

Validate

Run the lightweight layout check:

.\scripts\validate-layout.ps1

If you have the Codex skill creator validator available, also run:

python "$env:CODEX_HOME\skills\.system\skill-creator\scripts\quick_validate.py" ".\skills\spec"
python "$env:CODEX_HOME\skills\.system\skill-creator\scripts\quick_validate.py" ".\skills\ship"

License

MIT.