Open sourceMIT licence

Claude Code plugins

Two plugins from the rational-agents marketplace. Install either one on its own — into a single repository, or for yourself everywhere.

View on GitHub

01v0.1.0

adr-consensus

Drives an Architecture Decision Record to an implementation-readiness verdict — READY, NEEDS_DECISION or NOT_READY — through repeated independent review, scoped revision and consistency sweeps.

A review quorum of Anthropic models you map per project
Formats documents into the three-section ADR shape the review needs
Commits a recovery point before it rewrites the ADR; never pushes

How it works Source

02v0.1.0

replicants

Role-based engineering agents — implementer, analyst, QA, architect, deploy — and task skills for human-in-the-loop work. Each binds to the repository it runs in, so one plugin works in any project.

Seven agents and seven skills
Reads your CLAUDE.md and probes the repository before acting
Never invents a command, path, port or credential it did not find

How it works Source

Install

Both plugins, one marketplace

$claude plugin marketplace add rational-agents-ai/rational_agents_plugins --scope project$claude plugin install adr-consensus@rational-agents --scope project$claude plugin install replicants@rational-agents --scope project

Run from the repository you want the plugins in. Recorded in its .claude/settings.json, so everyone who works there gets the same setup.

Update or remove

claude plugin marketplace update rational-agentsclaude plugin uninstall adr-consensus@rational-agents --scope project

Plugin skills run under namespaced names, for example /adr-consensus:adr_consensus and /replicants:qa.