v1.0.0

Charter v1.0.0 — AI-agent readiness, scored

Charter gives any repository a single, deterministic readiness score for AI
coding agents — and shows you exactly what to fix. It runs entirely on your
machine: no LLM calls, no network, no telemetry.

Why it exists

Coding agents fail quietly on repos that lack context, leak secrets through
agent config, or wire up unpinned MCP servers. Charter catches those failure
modes before an agent does, scores them, and returns prioritized, fixable
findings — in your terminal, your CI, or a shareable report.

What you get

• charter doctor — score a repo against 18 agent-readiness rules
  (context, secrets, MCP safety, agent config, CI/env, testing/autonomy,
  governance), with a per-category scorecard and an interactive -i TUI.
• charter fix — diff-first repairs; nothing changes without your say-so.
• charter report — a self-contained HTML report (also Markdown/JSON) you
  can hand to a teammate or attach to a PR.
• charter explain · init · suppress · version round out the
  seven-command surface.
• CI-native — SARIF straight into GitHub Code Scanning, plus the official
  GitHub Action with a score gate that fails the build below your threshold.

Install

brew install use-charter/tap/charter
go install go.use-charter.dev/charter/cmd/charter@latest

Or download a signed binary for your platform below. Docs → https://use-charter.dev/docs

Verify what you run

Every release is cosign-signed with SLSA provenance and ships an SBOM per
artifact. Verify the checksums before you trust the binaries:

cosign verify-blob --bundle checksums.txt.sigstore.json checksums.txt

---

Apache-2.0 · offline-first · no LLM in the core · built in the open at
use-charter/charter.