jfmaes

jfmaes/awesome-ai-workflow

Plugin Claude Code

Plugin marketplace listing 1 plugin: ultimate-agentic-workflow.

jfmaes/awesome-ai-workflow

Plugin Claude Code

Tiered agentic workflow (T0-T3) with traceability, anti-slop gates, orchestration and context-engineering references, a preflight readiness check, and an installable .claude starter kit (subagents, stop-gate hooks, retro/mint-skill skills).

jfmaes/awesome-ai-workflow

Skill Claude CodeCodex

Routes AI coding work by risk tier (T0-T3) and enforces traceability, verification, and anti-slop discipline. Use when initializing or scaffolding a repo for AI coding (preflight readiness check, AGENTS.md/CLAUDE.md bootloaders, .claude starter kit with subagents/hooks/skills), migrating or upgrading a repo from a…

jfmaes/awesome-ai-workflow

Agent

Expert code review specialist for quality, security, and spec compliance. Use proactively after writing or modifying code, before claiming any T2+ task complete.

jfmaes/awesome-ai-workflow

Agent

Fresh-context implementer for one approved task from a plan. Use for subagent-driven development when a task has an exact spec, exact files, and a disjoint write scope.

jfmaes/awesome-ai-workflow

Agent

Read-only codebase and documentation researcher. Use for bounded discovery questions — how something works, where something lives, what patterns exist — when the exploration would flood the main context.

jfmaes/awesome-ai-workflow

Agent

Adversarial verifier that tries to refute claims of completion or findings. Use before accepting any "done" claim, any load-bearing finding, or any subagent report on T2+ work.

jfmaes/awesome-ai-workflow

Agent

Runs test suites, linters, typechecks, and builds, absorbing their verbose output. Use whenever checks must run and the full log would pollute the main context.

jfmaes/awesome-ai-workflow

Skill Claude CodeCodex

Use when deciding whether to adopt, migrate to, or reject a new tool, framework, model, library, or technique — triggered by a news brief, an HN/Twitter/Reddit release, a "should I switch to X?" or "is X better than what I use?" question, or FOMO about a trendy tool. Also when recording where a technology sits on a…

jfmaes/awesome-ai-workflow

Skill Claude CodeCodex

Create a new Claude Code skill or subagent from a repeated workflow, with trigger-description tuning and a quality checklist. Use when a workflow has been performed manually two or more times, when /retro proposes a skill candidate, or when the user asks to make something reusable.

jfmaes/awesome-ai-workflow

Skill Claude CodeCodex

Session retrospective that turns this session's friction and discoveries into durable improvements — OPS.md notes, new skill or subagent candidates, and tracked debt. Use at the end of substantive sessions, when the learning Stop hook fires, or when the user asks what was learned.

jfmaes/awesome-ai-workflow

Skill Claude CodeCodex

Use when a question needs a thorough, multi-source, fact-checked research report instead of a single-pass answer — evaluating a technology/protocol/vendor, "what is the current state of X", comparing options, verifying a claim before acting on it, or any research where sources should be cited and claims verified. For…