rosetta-preflight

rosetta-preflight is a command for Claude Code from tjboudreaux/rosetta. It costs 44 tokens per session (544 once invoked), scanned A, original, MIT.

A command that prepares a deterministic JSON preflight report for a scoped software change. It checks decision records, code scope, and local Rosetta rules before work is handed to an automated runner.

In plain words
What is it for?
Use it to review changed paths, decision coverage, and project records before starting a loop-runner task.
Why use it?
It makes missing evidence, unresolved decisions, and local gate failures visible before execution begins.

Command for Claude Code

Written for Claude Code: allowed-tools in frontmatter.

Part of the rosetta plugin — 1 skill, 3 commands shipped together

Install

Getting it into your agent

One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.

agentmods
npx agentmods add commands/tjboudreaux/rosetta/rosetta-preflight
Clone the repo
git clone --depth 1 https://github.com/tjboudreaux/rosetta

Made for: Claude Code.

Or install rosetta, the plugin that ships this one along with the rest of its 1 skill, 3 commands.

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.

agentmods badge for rosetta-preflight

README.md
[![agentmods](https://agentmods.dev/badge/commands/tjboudreaux/rosetta/rosetta-preflight.svg)](https://agentmods.dev/commands/tjboudreaux/rosetta/rosetta-preflight)
Your own site
<a href="https://agentmods.dev/commands/tjboudreaux/rosetta/rosetta-preflight"><img src="https://agentmods.dev/badge/commands/tjboudreaux/rosetta/rosetta-preflight.svg" alt="Measured on agentmods" height="20"></a>
Per session 44 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 544 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. 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.

ModelPer sessionOnce invoked
Fable 5.1 $0.00044 $0.00544
Opus 5 $0.00022 $0.00272
Sonnet 5 $0.00009 $0.00109
Haiku 4.5 $0.00004 $0.00054

Measured 6d ago against content hash 3240418559a8, method: parsed. Prices are Anthropic first-party input rates as of 2026-09-06, from the pricing page.

Security

Grade A, and why

rosetta-preflight scanned grade A with 0 findings 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 6d 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.

Nothing flagged

None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.

skills/rosetta/commands/rosetta-preflight.md · 38 lines

What it actually says

/rosetta-preflight — deterministic preflight JSON

Rosetta's deterministic CLI is local and does not call external APIs except rosetta preflight --allow-ra1-github, which delegates GitHub-dependent checks to RA1. Agent-run external-source collection for ADR 0012 is outside the deterministic CLI, opt-in, may use authenticated MCP/network tools, and may only feed rosetta ingest records as Status: Proposed drafts pending human confirmation. Rosetta is read-only against transcript stores and product source by default; default writes are limited to .agents/**, decisions/**, and loop-runs/**, plus the allowlisted harness docs only under explicit harness export --apply. Rosetta records, cites, and checks evidence; it never runs product builds/tests/deploys, asserts behavior, schedules loops, merges/pushes, or grades autonomy.

Workflow

Run the canonical top-level command; do not use a nested rosetta loop namespace:

rosetta preflight \
  --project-root <project> \
  --decisions-root <decisions-root> \
  --scope "<decision-query>" \
  --min-coverage 0.8 \
  [--changed-path <path> ... | --diff-file <path-or-> | --base <rev> --head <rev>] \
  [--change-id <stable-id>] \
  [--allow-ra1-github]

Interpret the output sections by owner:

  • ra1_structural (owner: ra1) — structural readiness JSON from RA1, or skip when RA1 is absent.
  • decision_state (owner: rosetta) — current decision resolution for --scope; fails on conflict, ambiguous alias, stale current records, or non-unique matches.
  • gates (owner: rosetta) — the same pure gate report as rosetta gates check.

Exit 1 means at least one section failed. Exit 0 means the deterministic preflight report was generated with no failing sections; execution policy remains outside Rosetta.

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.

  1. 6d ago First seen · 38 lines · 0 tokens per session scan A 3240418559a8

Subscribe to this mod's changes

rosetta-preflight is a command published in the GitHub repository tjboudreaux/rosetta (2 stars, last pushed 1mo ago), licensed MIT. It adds 44 tokens to every session and 544 once invoked, about $0.0002 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.