A GitHub Copilot CLI plugin providing essential GitHub delivery and maintenance workflows — agents and skills for CI/CD, releases, security, incidents, and team operations.
copilot plugin install https://github.com/kellyirvin/copilot-cli-essentials
Agents are specialized AI assistants invoked with @agent-name. Each has a defined role, curated tool access, and pre-loaded skills.
| Agent |
Description |
@release-engineer |
Orchestrates end-to-end releases — versioning, changelog, tagging, GitHub release, and post-release validation |
@security-auditor |
Security-focused code review and vulnerability assessment before merges and releases |
@incident-commander |
Incident response coordination from triage through mitigation and post-mortem |
@code-reviewer |
Comprehensive PR review covering correctness, security, DX, CI/CD, and CODEOWNERS |
@devils-advocate |
Poke holes in plans, deployments, and architecture decisions before committing |
@log-reader |
Efficiently triage build and deployment failures across GitHub Actions, CircleCI, and Jenkins |
@haiku |
Lightweight executor for well-defined delegated tasks |
Skills are on-demand knowledge and workflow documents. They are loaded automatically by agents when relevant.
| Skill |
Description |
deploy |
Run the deployment workflow — pre-flight checks, environment targeting, production gate, and rollback |
release |
Cut a versioned release — semver bump, changelog, tag, and GitHub release |
incident |
Incident response workflow — triage, investigate, mitigate, communicate, and post-mortem |
workflow |
Debug and fix CI/CD pipeline failures |
standup |
Generate standup notes from git history, PRs, and issues |
sync |
Sync a fork or branch with upstream |
onboard |
Orient to a new project — structure, conventions, CI/CD, and workflows |
| Skill |
Description |
review |
Comprehensive code review of the current branch |
audit |
Security and dependency audit — vulnerabilities, secrets, outdated packages |
fix-issue |
Fetch a GitHub issue and implement a fix |
test |
Run tests and analyze failures |
commit |
Stage changes, run preflight checks, and create a conventional commit |
pr |
Create a pull request with auto-generated title and description |
deps |
Audit, check, or upgrade project dependencies |
| Skill |
Description |
explain |
Deep explanation of a file, function, or concept |
execute |
Find and execute an implementation plan from the plans folder |
init |
Initialize GitHub repo best practices — CODEOWNERS, Dependabot, branch protection, PR templates |
| Skill |
Description |
ci-cd-pipelines |
GitHub Actions patterns, caching, security, matrix builds |
release-management |
Semver, changelog authoring, release branching |
security-scanning |
OWASP Top 10, secret detection, dependency CVEs |
dependency-management |
CVE triage, safe upgrades, Dependabot/Renovate configuration |
branching-strategy |
Trunk-based development, naming conventions, PR workflow |
incident-response |
Severity triage, investigation playbooks, post-mortems |
deployment-strategies |
Blue/green, canary, feature flags, database migrations |
code-ownership |
CODEOWNERS setup, review standards, team routing |
reading-logs |
Log investigation across GitHub Actions, CircleCI, and Jenkins |
| Skill |
Description |
architecting-systems |
Clean system architecture, module boundaries, dependency management |
writing-tests |
Behavior-focused tests using the Testing Trophy model |
refactoring-code |
Structure improvements while preserving behavior |
systematic-debugging |
Four-phase root cause analysis before proposing fixes |
handling-errors |
Preventing silent failures and context loss |
migrating-code |
Safe migrations with backward compatibility and reversibility |
optimizing-performance |
Measure-first performance work |
fixing-flaky-tests |
Diagnosing tests that fail when run concurrently |
condition-based-waiting |
Replacing arbitrary timeouts with condition polling |
managing-databases |
Schema design, query optimization, PostgreSQL/DuckDB/PGVector |
preflight-checks |
Detect and run linters, formatters, and type checkers |
verification-before-completion |
Verification standards before claiming work is done |
post-mortem |
Session review to extract actionable improvements |
planning-products |
Feature scoping, product specs, JTBD analysis |
writing-plans |
Implementation plans grouped by subsystem |
executing-plans |
Orchestrating multi-agent plan execution |
| Skill |
Description |
writer |
Writing style guide for human-sounding content |
documentation |
Task-oriented technical documentation |
design |
Precise, minimal UI design for dashboards and admin interfaces |
visualizing-with-mermaid |
Professional Mermaid diagrams with semantic styling |
configuring-copilot |
How to write skills, agents, and hooks for this plugin system |
MIT