8000 GitHub - rezkam/boring-but-good: A personal collection of AI agent skills that are boring, but good. Monorepo of custom Agent Skills. · GitHub
[go: up one dir, main page]

Skip to content

About

A personal collection of AI agent skills that are boring, but good. Monorepo of custom Agent Skills.

Resources

Stars

1 star

Watchers

0 watching

Forks

Latest commit

 

History

217 Commits

Folders and files

Repository files navigation

boring-but-good

I was bored. So I wrote some scripts to make my AI agents actually useful at the tedious stuff I don't want to do — checking builds, filing tickets, reviewing vulnerabilities, that kind of thing. Called it "boring" because that's what this work is. Called it "good" because it actually works.

What's in here

Skills that give coding agents the ability to interact with real engineering systems through shell scripts and workflow instructions. Publishable skills live under skills/<skill-name>/, matching the standard skills repository layout.

Skill What it does
jira Create, view, update, transition, and search Jira issues. Works with Cloud and Server/DC via go-jira.
to-tasks Break plans/specs into proposed tasks, ask whether to create them in Jira or locally, then create them after approval.
jenkins Check build status, read test failures, view console output, trigger builds, watch pipelines.
sonarqube Fetch code quality issues, coverage metrics, security hotspots, quality gate status.
dependency-track Query SCA findings, audit vulnerabilities, check project health, review policy violations.
argocd Check app sync status, trigger syncs, view resources, read pod logs, rollback deployments.
skanetrafiken Plan public transport journeys in southern Sweden with real-time delays.
java-21-to-25-migration Migrate a Java project from JDK 21 to JDK 25 with a phased plan covering all breaking changes.
finance Fetch browser-authenticated market and macroeconomic data through Browser Tools managed Chrome.
ai-chat Query browser-authenticated AI providers and power the file-backed gemini_search pi extension.
code-review Review PRs, branches, local changes, snapshots, or codebases for concrete, material defects without editing code.
commit Stage and commit only the intended changes.
handoff Compact the current conversation into a handoff document for the next agent.
perplexity Search, research, chat, and analyze files through Perplexity WebUI.
pr-ready Commit, push, open, and verify a merge-ready pull request without merging it.
research Research a question against primary sources in a background agent and save cited findings as Markdown.
ship Verify, then open a pull request and watch CI, or push straight to main when the repository ships that way.
tdd Write a failing test first, prove it detects the behavior, then make it pass.
verify Establish that a change or the whole codebase works by formal proof and reproducible executable verification, and gate release with PASS, FAIL or BLOCKED.

Pi extensions live separately under extensions/. This keeps executable harness integrations independent from the skills that document their workflows.

Browser Tools is an external companion for managed Chrome automation. Install its package with npm install @rezkam/browser-tools, or install its skill from the standalone repository with npx skills add rezkam/browser-tools.

Getting started

Quick install (any agent)

Install skills to any supported agent (Claude Code, Cursor, Codex, etc.):

# Install all skills
npx skills add your-org/boring-but-good

# Install a specific skill
npx skills add your-org/boring-but-good --skill jira

# Install to a specific agent
npx skills add your-org/boring-but-good --skill jenkins -a claude-code

# List available skills without installing
npx skills add your-org/boring-but-good --list

Manual install

git clone https://github.com/your-org/boring-but-good.git
cd boring-but-good
./setup.sh

The setup script walks you through configuring whichever skills you need. It creates symlinks from the repo into your agent's skill directory, so git pull updates everything in place.

If you installed from a version that kept skills at the repository root or under workflow-tools/, run ./setup.sh once after updating to refresh the existing symlinks.

How it works

Script-backed skills follow this structure:

skills/skill-name/
├── SKILL.md        # Agent reads this to know what's available
├── scripts/
│   ├── _config.sh  # Loads credentials from ~/.boring/<skill>/
│   ├── _api.sh     # HTTP helper (one place for all curl calls)
│   └── *.sh        # One script per operation
└── README.md       # You're reading the human version

Credentials live in ~/.boring/<skill>/ as separate files (url, token, etc.). Never in the scripts, never in the repo.

Tests

./tests/test-all.sh

Covers argument validation, error messages, API compatibility, URL encoding, pagination, and regression cases for every bug we've fixed.

License

Apache License 2.0 — see LICENSE.

About

A personal collection of AI agent skills that are boring, but good. Monorepo of custom Agent Skills.

Resources

Stars

1 star

Watchers

0 watching

Forks

Releases

Packages

Contributors

Languages

0