AI-agent readiness, scored.
AI agents are only as good as the repo they work in — missing context, exposed secrets, and unpinned tools quietly wreck their output. Charter grades yours in under two seconds: a 0–100 readiness score across nine categories, every gap paired with a concrete fix. No cloud, no account, no telemetry — it runs fully offline, and the same check holds the line in CI.
brew install use-charter/tap/charter 
Claude Code 
Cursor 
Copilot 
Windsurf 
Codex 
Gemini 
Replit 
Zed Nine categories. One honest score.
Agents trip on the same handful of things — missing context, secrets in plain sight, an unpinned MCP server, no way to verify their own work. This is the real v1 catalog: every rule carries an ID, a severity that sets its score weight, and the exact thing it checks.
A small, sharp command surface
No dashboard. No account. No telemetry. Seven commands that behave identically in your shell and in CI. Click any one to watch it run.
Scan the repo and print a 0–100 readiness score with a per-category breakdown. Exit 0 pass, 1 below threshold, 2 error.
Scaffold the context files an agent needs, so the next scan starts from a clean baseline. Never overwrites.
Diff-first auto-repair for supported rules. Nothing is written until you approve the diff.
Record a governed waiver — with a reason, an owner, and an expiry — instead of silently ignoring a rule.
Print the catalog metadata for any rule: id, name, category, summary, and docs URL.
Write a self-contained, offline single-file HTML report — fonts and data inlined, opens from a file URL.
Print the version, build provenance, and supply-chain verification status. Useful in CI logs.
The score only ratchets up
Readiness isn't a one-time cleanup. Charter walks a repo from an honest baseline to ship-ready, then turns the score into a wall every pull request has to clear.
Make the score a merge requirement
Add one workflow and Charter re-scores every pull request, failing the check when the score drops below your threshold. Findings post as SARIF 2.1.0, so they land inline in GitHub Code Scanning — no new dashboard to learn.
No cloud. No account. No catch.
Charter runs entirely on your machine and shows its work. The whole contract is a handful of flags you can verify yourself — right down to the signature on the binary.
Nothing ever leaves your machine.
No network, no LLM calls, no telemetry. Charter reads your files locally and prints a number. That's the entire deal.
Free, forever
Open source and permissively licensed. Vendor it, fork it, ship it inside your own tooling — no seat count, no expiry.
Supply-chain signed
Every release ships SLSA Level 3 provenance, cosign-signed. Verify the exact binary you install.
Standards in, standards out
SARIF 2.1.0 drops straight into GitHub Code Scanning and the tools you already run.