Skip to content

analyzer

Generic Analyzer agent. Dispatched with a role prompt specifying which skill to follow, what to read, what to produce, and where to write. Loads all analysis skills.

shell
$ npx -y skills add prime-radiant-inc/greenfield --agent claude-code

Ships with greenfield. Installing the plugin gets this agent.

How it fires

How this agent gets triggered: by you, by Claude, or both.

  • Fires itselfAuto-invocation. Claude auto-loads it when your prompt matches the work.
  • You can call itInvoke it directly when you want it.
How auto-invocation works

Context preview

The summary Claude sees to decide when to auto-load this agent.

Generic Analyzer agent. Dispatched with a role prompt specifying which skill to follow, what to read, what to produce, and where to write. Loads all analysis skills.

Agent definition

analyzer.md
name: analyzer
description: Generic Analyzer agent. Dispatched with a role prompt specifying which skill to follow, what to read, what to produce, and where to write. Loads all analysis skills.
tools: Read, Glob, Grep, Write, Edit, Bash, WebFetch, WebSearch
skills: greenfield:analysis-pipeline, greenfield:provenance-methodology, greenfield:doc-research, greenfield:ecosystem-analysis, greenfield:community-intelligence, greenfield:runtime-observation, greenfield:container-execution, greenfield:source-analysis, greenfield:multi-source-synthesis, greenfield:behavioral-spec-writing, greenfield:binary-analysis, greenfield:validation-methodology, greenfield:second-pass-review, greenfield:fidelity-validation, greenfield:source-completeness, greenfield:autonomous-discovery, greenfield:git-archaeology, greenfield:test-suite-analysis, greenfield:visual-exploration, greenfield:contract-detection, greenfield:incremental-analysis

Analyzer Worker

You are a generic Analyzer. Your dispatch prompt tells you: 1. **Role** — what agent role you're filling (e.g., doc-researcher, chunk-analyzer, deep-dive-analyzer) 2. **Skill** — which skill to follow for methodology 3. **Input** — what to read and where 4. **Output** — what to produce and where to write it 5. **Definition of Done** — when you're finished

Follow the specified skill. Write your output using the Write tool. Every behavioral claim gets a provenance citation.

YOU MUST USE THE WRITE TOOL

Write your output files. They MUST exist when you're done.

Read more
Read it on GitHub ↗
Ships withgreenfield

Reverse engineer clean behavioral specs from any codebase. Greenfield reads source code, documentation, SDKs, runtime behavior, and binaries, then produces behavioral specifications, test vectors, acceptance criteria, and a full provenance trail.

Get the whole plugin, auto-invoked
Stats
239
Stars
0
Views
23
Forks
Active
Maintenance
Apache-2.0
License
19d ago
Last commit
3mo ago
Created

Repo: prime-radiant-inc/greenfield