agentic-engineering

Plugin agentic-engineering from marketplace diamantai in NirDiamant/Agentic_Engineering. Commands read from the repo's manifest — 16 days ago.

Skills (5)

• spark

  The human on-ramp to the agent-harness. Turn an idea that only exists in your head into a running build, without writing a spec yourself. Leads you through a short interview about the ESSENCE of what you're making (the purpose, the image in your mind, the core behavior), then distills it into a buil

• docs-gardener

  Autonomously capture the *why* behind a work session into a project's docs: the human decisions, tradeoffs, connections, and steering points that can't be inferred from the code. Runs at the end of a session (or on demand), records ONLY what was actually decided or discovered this session, appends d

• start-project

  The technical on-ramp to a fresh project. Where /spark draws out the ESSENCE and refuses to talk tech, start-project does the opposite. It surveys the technical infrastructure up front (existing setup, language, database, cloud/hosting, CI/CD, deployment, testing, package manager), turns the answers

• tester

  Try to fail what just changed. The scoped, everyday complement to /testify. Reads the conversation or PR diff, picks the few behaviors most worth attacking (the art of covering enough without over-testing), writes tests that attempt to break them on the existing runner, summarizes each attempt with

• testify

  Stand up a project's tests and prove they're worth anything. Where /start-project writes docs/TESTING.md (the intent, meaning the bar, the layers, the runner) but doesn't touch code, testify does the doing. It detects the stack, wires the runner, writes real tests across the layers (unit, integratio

Slash commands (2)

• /agentic-engineering:agent-memory

  Give this repo a memory. Reads the real code, writes CLAUDE.md and a small docs layer from it, grades the repo 0 to 5 before and after, lists the claims in your own docs the code contradicts. Deletes nothing, overwrites nothing, stops once for approval. About fifteen minutes.

• /agentic-engineering:org-card

  One card for a whole team. Points at a folder of repositories, or at a shared monorepo with --owner to read only that team's folders, reads each one cold (the same five questions the single-repo check asks), and prints one org card: how many repos score 2 or below, what your agents guess about most,

Similar plugins