Complex product work can require coordination across planning, coding, testing, and operations. Loki Mode provides a structured multi-agent workflow with task queues, checkpoints, and review steps.

LLM applications can fail without clear patterns for retrieval, agents, prompts, and monitoring. This skill gives practical templates for RAG, agents, prompt workflows, and LLMOps.

Weak prompts create inconsistent LLM behavior, higher token costs, and difficult testing. This skill turns rough instructions into structured prompts with evaluation, safety, and deployment guidance.

Production LangChain agents need clear architecture, memory, evaluation, and observability. This skill guides Claude, Codex, and Claude Code through LangGraph patterns, async tools, RAG, testing, and deployment.

Teams need reliable guidance for building conversational AI systems. This skill provides architecture, testing, deployment, and monitoring patterns for production assistants.

Linux incidents often start with broad symptoms and scattered evidence. This skill provides a phased workflow for collecting facts, analyzing resources, checking logs, testing networking, and validating fixes.

Linux administrators need repeatable scripts for backups, monitoring, and maintenance. This skill provides Bash templates and guidance for common operational tasks.

Linux teams need repeatable scripts for common administration work. This skill provides structured Bash templates for backups, monitoring, logs, users, networking, and automation.

Linux privilege escalation reviews require structured checks and careful authorization. This skill organizes enumeration, exploit-path analysis, evidence capture, and remediation for controlled security assessments.

Code changes often fail because lint, type, or audit checks are skipped. This skill guides agents through repeatable validation for Node.js and Python projects.

LinkedIn API workflows are hard to sequence because IDs, OAuth scopes, and image uploads must be handled correctly. This skill guides Claude, Codex, and Claude Code through common LinkedIn workflows using Rube MCP.

Teams lose time when issue updates, project links, and status changes are handled manually. This skill gives Claude, Codex, and Claude Code structured Linear workflows using MCP, CLI, and helper-script patterns.

Teams lose time on repetitive Linear issue, project, and cycle tasks. This skill guides Claude, Codex, and Claude Code through Rube MCP Linear workflows.

Teams need clear legal templates for software products and data practices. This skill drafts structured documents and flags areas for attorney review.

Legacy systems are hard to change without breaking users or integrations. This skill guides incremental upgrades, compatibility planning, tests, and rollback paths.

Launches often stall when teams rely on one announcement. This skill helps plan phased campaigns, channel strategy, Product Hunt launches, and post-launch momentum.

Fast-changing topics are hard to research from memory. This skill gathers recent Reddit, X, and web signals, then turns them into practical guidance.

Laravel applications can hide authorization, validation, and deployment weaknesses. This skill guides Claude, Codex, and Claude Code through structured security reviews.

Laravel projects need consistent architecture, validation, security, and performance discipline. This skill guides Claude, Codex, and Claude Code toward idiomatic Laravel solutions.

Stateful agent systems can become hard to trace and maintain. This skill helps Claude, Codex, and Claude Code design clearer LangGraph workflows.