A command that coordinates an optional prototype for the current software specification. A prototype is an early trial used to test an idea before full implementation.

Use it to start, resume, suggest, or cancel a prototype and to manage its review, verdict, handoff, and cleanup steps.

It lets a team investigate an uncertain approach while keeping the main specification phase and checkout unchanged. It also records whether the prototype should be resumed, canceled, or handed off.

prototype — Command for Claude Code by tzachbon: A command that coordinates an optional prototype for the…

Install

Getting it into your agent

This one installs as part of its plugin. Adding the marketplace and installing the plugin brings it with everything else the plugin ships.

Claude Code

/plugin marketplace add tzachbon/smart-ralph

Claude Code

/plugin install ralph-specum

Made for: Claude Code.

Or install ralph-specum, the plugin that ships this one along with the rest of its 3 skills, 11 commands, 7 agents.

Wrote this? Show the measurements

A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.

README.md

[![agentmods](https://agentmods.dev/badge/commands/tzachbon/smart-ralph/prototype.svg)](https://agentmods.dev/commands/tzachbon/smart-ralph/prototype)

Your own site

<a href="https://agentmods.dev/commands/tzachbon/smart-ralph/prototype"><img src="https://agentmods.dev/badge/commands/tzachbon/smart-ralph/prototype.svg" alt="Measured on agentmods" height="20"></a>

Per session 6 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.

When invoked 423 The whole file, excluding the scripts and references it only reads on demand.

Security scan B 1 finding. Scan, not verified.

Origin original No closer match found in the catalogue.

Token cost

What it costs to keep this loaded

Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.

Model	Per session	Once invoked
Fable 5.1	$0.00006	$0.00423
Opus 5	$0.00003	$0.00211
Sonnet 5	$0.00001	$0.00085
Haiku 4.5	$0.00001	$0.00042

Measured 2d ago against content hash c52299a0c1fd, method: parsed. Prices are Anthropic first-party input rates as of 2026-09-06, from the pricing page.

Security

Grade B, and why

prototype scanned grade B with 1 finding against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 2d ago.

A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.

Unrestricted tool accessmediumExcessive agency

A wildcard tool grant or "run any command" leaves no least-privilege boundary at all.

allowed-tools: "*"

plugins/ralph-specum/commands/prototype.md · 22 lines

What it actually says

Prototype

Coordinate one optional prototype overlay. Do not implement source work in this command.

Parse $ARGUMENTS. Accept direct invocation by default, --suggested --return-phase <phase>, --resume <id>, --cancel <id>, or --quick. A spec name may precede the mode.
Source ${CLAUDE_PLUGIN_ROOT}/hooks/scripts/path-resolver.sh and call ralph_resolve_context. If a spec name was supplied, replace the context basePath with the verified result from ralph_find_spec and set specRoot to its parent. Stop if basePath is null or missing.
Read ${CLAUDE_PLUGIN_ROOT}/references/prototype-coordinator.md completely and follow it without reimplementing its mechanics.
Use only the resolved basePath and these helpers:
- ${CLAUDE_PLUGIN_ROOT}/hooks/scripts/locked-state.py
- ${CLAUDE_PLUGIN_ROOT}/hooks/scripts/prototype-records.py
- ${CLAUDE_PLUGIN_ROOT}/hooks/scripts/prototype-harness.py
Preserve the current main phase and current checkout while the overlay runs. Stop at safe tool boundaries. Require the exact normal-mode approvals for capture, dirty paths, verdict, handoff, deletion, and every remote action.
Store and resume the request in activePrototypes with returnPhase and any returnTaskIndex. Reconcile immutable records before dispatch or publication.
In quick mode, let the agent own all choices, keep the run bounded, review any post-deletion bytes, publish one terminal outcome, and continue to design without a user stop.
Never push source or evidence, update a remote issue, or publish a branch without separate explicit authorization.

Changes

What this file has done since we first saw it

Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.

2d ago First seen · 22 lines · 6 tokens per session scan B c52299a0c1fd

Subscribe to this mod's changes

prototype is a command published in the GitHub repository tzachbon/smart-ralph (533 stars, last pushed 3d ago), licensed MIT. It adds 6 tokens to every session and 423 once invoked, about $0.0000 per session on Opus 5. A static security scan graded it B with 1 finding (unrestricted tool access). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-09-03.

Planning agentic-coding ai ai-agents autonomous-coding claude claude-code

prototype

Getting it into your agent

Wrote this? Show the measurements

What it costs to keep this loaded

Grade B, and why

Unrestricted tool accessmediumExcessive agency

What it actually says

Prototype

What this file has done since we first saw it

Other commands, from other repositories

psm

project

version

discover

diff-docs

voice-compliance