~ / rigs / michaelmarconi / agentic-engineering

michaelmarconi/agentic-engineering

A software development lifecycle for Claude Code: specs, decision records and plans, independently reviewed, taking work from idea to verified code.

↗ GitHub ★ 0 MIT updated today project Claude CodeCodex Claude plugin
stack: Python
share on X
ARCHETYPE
Orchestrator
A bench of specialised subagents. The main agent mostly delegates.
CONTEXT TAX · EVERY TURN
~7.7k tokens
Heavy · median rig: 2.3k · breakdown
GUARDRAILS
0/5
No committed guardrails · details
Quality 5/100 Guardrails v2

Copy this rig

# review before running: this installs third-party code
$ npx degit michaelmarconi/agentic-engineering/agents ./rig-agentic-engineering/agents
$ npx degit michaelmarconi/agentic-engineering/skills ./rig-agentic-engineering/skills
$ npx degit michaelmarconi/agentic-engineering/hooks ./rig-agentic-engineering/hooks

MCP servers are added to Claude Code at local scope; env vars are shown as YOUR_… placeholders — we never store values. Files are fetched with degit into a separate folder so you can review before merging.

$ npx degit michaelmarconi/agentic-engineering/skills/adr .claude/skills/adr
$ npx degit michaelmarconi/agentic-engineering/skills/audit .claude/skills/audit
$ npx degit michaelmarconi/agentic-engineering/skills/brief .claude/skills/brief
$ npx degit michaelmarconi/agentic-engineering/skills/feedback .claude/skills/feedback
$ npx degit michaelmarconi/agentic-engineering/skills/plan .claude/skills/plan
$ npx degit michaelmarconi/agentic-engineering/skills/practices .claude/skills/practices
$ npx degit michaelmarconi/agentic-engineering/skills/release .claude/skills/release
$ npx degit michaelmarconi/agentic-engineering/skills/spec .claude/skills/spec
$ npx degit michaelmarconi/agentic-engineering/skills/spike .claude/skills/spike
$ npx degit michaelmarconi/agentic-engineering/skills/verify .claude/skills/verify
$ curl -fsSL --create-dirs -o .claude/agents/adr-reviewer.md https://raw.githubusercontent.com/michaelmarconi/agentic-engineering/main/agents/adr-reviewer.md
$ curl -fsSL --create-dirs -o .claude/agents/audit-checker.md https://raw.githubusercontent.com/michaelmarconi/agentic-engineering/main/agents/audit-checker.md
$ curl -fsSL --create-dirs -o .claude/agents/audit-picker.md https://raw.githubusercontent.com/michaelmarconi/agentic-engineering/main/agents/audit-picker.md
$ curl -fsSL --create-dirs -o .claude/agents/auditor.md https://raw.githubusercontent.com/michaelmarconi/agentic-engineering/main/agents/auditor.md
$ curl -fsSL --create-dirs -o .claude/agents/brief-checker.md https://raw.githubusercontent.com/michaelmarconi/agentic-engineering/main/agents/brief-checker.md
$ curl -fsSL --create-dirs -o .claude/agents/brief-reviewer.md https://raw.githubusercontent.com/michaelmarconi/agentic-engineering/main/agents/brief-reviewer.md
$ curl -fsSL --create-dirs -o .claude/agents/build-correction-checker.md https://raw.githubusercontent.com/michaelmarconi/agentic-engineering/main/agents/build-correction-checker.md
$ curl -fsSL --create-dirs -o .claude/agents/build-reviewer.md https://raw.githubusercontent.com/michaelmarconi/agentic-engineering/main/agents/build-reviewer.md
$ curl -fsSL --create-dirs -o .claude/agents/correction-checker.md https://raw.githubusercontent.com/michaelmarconi/agentic-engineering/main/agents/correction-checker.md
$ curl -fsSL --create-dirs -o .claude/agents/feedback-reviewer.md https://raw.githubusercontent.com/michaelmarconi/agentic-engineering/main/agents/feedback-reviewer.md
$ curl -fsSL --create-dirs -o .claude/agents/language-checker.md https://raw.githubusercontent.com/michaelmarconi/agentic-engineering/main/agents/language-checker.md
$ curl -fsSL --create-dirs -o .claude/agents/plan-reviewer.md https://raw.githubusercontent.com/michaelmarconi/agentic-engineering/main/agents/plan-reviewer.md
$ curl -fsSL --create-dirs -o .claude/agents/practices-reviewer.md https://raw.githubusercontent.com/michaelmarconi/agentic-engineering/main/agents/practices-reviewer.md
$ curl -fsSL --create-dirs -o .claude/agents/release-reviewer.md https://raw.githubusercontent.com/michaelmarconi/agentic-engineering/main/agents/release-reviewer.md
$ curl -fsSL --create-dirs -o .claude/agents/spec-reviewer.md https://raw.githubusercontent.com/michaelmarconi/agentic-engineering/main/agents/spec-reviewer.md
$ curl -fsSL --create-dirs -o .claude/agents/spike-reviewer.md https://raw.githubusercontent.com/michaelmarconi/agentic-engineering/main/agents/spike-reviewer.md
$ curl -fsSL --create-dirs -o .claude/agents/verifier.md https://raw.githubusercontent.com/michaelmarconi/agentic-engineering/main/agents/verifier.md
$ curl -fsSL --create-dirs -o .claude/agents/verify-checker.md https://raw.githubusercontent.com/michaelmarconi/agentic-engineering/main/agents/verify-checker.md

This rig commits no guardrails. Here is the community baseline instead — the deny/ask rules most often found across all 16.2k rigs:

{
  "permissions": {
    "deny": [
      "Read(./.env)",
      "Read(./.env.*)",
      "Read(**/.env)",
      "Read(~/.ssh/**)",
      "Bash(git push --force:*)",
      "Read(**/*.pem)",
      "Read(.env)",
      "Read(~/.aws/**)",
      "Read(**/*.key)",
      "Read(**/.env.*)",
      "Bash(git push --force*)",
      "Bash(rm -rf:*)",
      "Bash(rm -rf *)",
      "Bash(sudo *)",
      "Bash(sudo:*)",
      "Bash(rm -rf /)",
      "Bash(git push -f:*)",
      "Read(.env.*)",
      "Bash(git reset --hard*)",
      "Bash(rm -rf /*)"
    ],
    "ask": [
      "Bash(git push:*)",
      "Bash(git push *)",
      "Bash(gh pr merge *)",
      "Bash(git commit:*)",
      "Bash(rm *)",
      "Bash(gh pr merge:*)",
      "Bash(gh pr create *)",
      "Bash(git reset --hard:*)",
      "Bash(npm publish:*)",
      "Bash(git rebase *)"
    ]
  }
}

Skills (10)

Subagents (18)

adr-reviewer
model: sonnet
Reviews one ADR (architecture decision record) against the ADR rubric and reports pass or fail for each check, with evidence. Started by the adr skill with `Document: <path>` as its request's first li
audit-checker
model: sonnet
Tries to disprove one auditor's drift findings about one accepted spec or ADR, rereading the code, or the brief checker's contradictions about it, rereading the brief, and says of each whether it is c
audit-picker
model: sonnet
Reads the files a window of commits changed beside every accepted spec and ADR, and names the ones those changes could affect. Started by the audit skill with `Repository: <folder>` as its request's f
auditor
model: sonnet
Checks the code against one accepted spec or ADR and reports only where they disagree, with quotes of both, plus how many requirements agree and which are not built. Started by the audit skill with `D
brief-checker
model: sonnet
Reads the product brief beside accepted specs and ADRs and reports only where one contradicts the brief's Goals, Out of scope or Constraints, with quotes of both. Started by the audit skill with `Docu
brief-reviewer
model: sonnet
Reviews the product brief against the brief rubric and reports pass or fail for each check, with evidence. Started by the brief skill with `Document: <path>` as its request's first line. To review a b
build-correction-checker
model: sonnet
Checks the corrections made to a build after its three reviews. Started by the build skill with `Document: <plan path>` as its request's first line; a plugin hook attaches the review reports and the c
build-reviewer
model: sonnet
Reviews one build against its plan and the build rubric, and reports pass or fail for each check, with evidence. Started by the build skill with `Document: <plan path>` as its request's first line. To
correction-checker
model: sonnet
Checks the corrections made to one document after its review, for every skill that writes documents. Started with the same `Document:` line the reviewer was given; a plugin hook attaches the review an
feedback-reviewer
model: sonnet
Reviews a draft issue for the agentic plugin against the feedback rubric, chiefly that it leaks nothing from the person's repository, and reports pass or fail for each check with evidence. Started by
language-checker
model: sonnet
Checks one document for words a newcomer could not follow, by the lifecycle's Plain words rule, and reports pass or fail with evidence. Started beside a skill's reviewer, with the same `Document:` lin
plan-reviewer
model: sonnet
Reviews one plan against the plan rubric and reports pass or fail for each check, with evidence. Started by the plan skill with `Document: <path>` as its request's first line. To review a plan, use th
practices-reviewer
model: sonnet
Reviews a project's recorded practices, the Practices section of AGENTS.md and any topic files under docs/practices/, against the practices rubric and reports pass or fail for each check, with evidenc
release-reviewer
model: sonnet
Reviews a draft of one release's notes against the release notes rubric, chiefly that every gathered note is covered and nothing is invented, and reports pass or fail for each check with evidence. Sta
spec-reviewer
model: sonnet
Reviews one spec against the spec rubric and reports pass or fail for each check, with evidence. Started by the spec skill with `Document: <path>` as its request's first line. To review a spec, use th
spike-reviewer
model: sonnet
Reviews one spike record against the spike rubric, its design or its verdict, and reports pass or fail for each check, with evidence. Started by the spike skill with `Document: <path>` as its request'
verifier
model: opus
Starts the product and uses it from outside, as a person would, to check each example of the named requirements of one accepted spec, and reports pass, fail or not checked for each, with evidence, nev
verify-checker
model: opus
Tries to disprove one verifier's verdicts about one accepted spec by using the running product again from outside, never reading the code, and says of each whether it is confirmed or rejected, with ev

Similar rigs

copied ✓