Engineering
swe-cli-skills avatar

swe-cli-skills

Expert CLI guides for AI agents, featuring senior engineer workflows, safety guardrails, and operational patterns for cloud, IaC, containers, databases, and dev tools.

Introduction

This skill acts as a comprehensive repository of senior-level CLI knowledge designed specifically for autonomous software engineering agents. Rather than providing standard reference documentation, it encapsulates operational judgment, safety guardrails, error recovery strategies, and best-practice composition patterns for 23 essential command-line tools. Targeted at professional developers and AI-powered coding assistants, this tool ensures that agents execute complex infrastructure and development tasks with the caution and precision of a seasoned engineer.

  • Covers critical operations across 9 domains: Cloud (AWS, gcloud, Azure), IaC (Terraform), Containers (Docker, kubectl, Helm), Git/VCS, Dev Tools (jq, sed, make), Networking (curl, SSH), Package Managers (npm, pip/uv), Databases (psql, redis-cli), and Platforms (Stripe, Sentry, Vercel, Firebase).

  • Provides expert-curated 'gotchas' and non-interactive workflow adjustments to prevent common failures in headless environments, such as terminal TTY hang-ups or production-blocking command errors.

  • Focuses on multi-tool composition, demonstrating how to pipe data effectively between utilities like AWS CLI, jq, and xargs for complex automation pipelines.

  • Offers localized, environment-aware configuration advice, highlighting portability differences between Linux and macOS environments for tools like sed and package managers.

  • Intended for use when an agent must perform high-stakes operations such as infrastructure provisioning, database migration, or CI/CD debugging.

  • Inputs should be specific operational queries or error messages; outputs provide direct command sequences, flag configurations, and safety verification steps.

  • Constraints include a strong emphasis on production safety; agents are guided to use non-destructive flags, explicit project context settings, and verification routines like terraform plan or dry-run modes before applying changes.

  • Effectively reduces hallucinations in agents by providing a deterministic, judgment-based reference for complex tool syntax and sequence logic.

Repository Stats

Stars
12
Forks
2
Open Issues
0
Language
Shell
Default Branch
main
Sync Status
Idle
Last Synced
May 3, 2026, 11:06 PM
View on GitHub