OpenCode already loads SKILL.md files through its native skill tool. Give it one that stops overpaying.

How this maps onto OpenCode

Same rubric. OpenCode's native skill tool underneath it.

OpenCode ships a documented Agent Skills mechanism: a native skill tool that discovers SKILL.md files under .opencode/skills/<name>/SKILL.md (walking up from your working directory to the git worktree root), and — for compatibility with skills written for other agents — the same <name>/SKILL.md layout under .claude/skills/ and .agents/skills/ too, plus the equivalent global paths under ~/.config/opencode/, ~/.claude/, and ~/.agents/. The agent calls skill({ name }) to load a skill's full content on demand, gated by the frontmatter's required name and description fields. That's the exact shape Undercut ships in — no adapter needed.

Undercut's tier names — cheap, standard, frontier, apex — are generic on purpose. models.md resolves each one to a real model slug per provider family (Anthropic, OpenAI, Google, open-weight), which lines up with how OpenCode itself works: it's provider-agnostic, so whichever model family you've wired OpenCode up to, the tier maps onto that family's roster the same way it does for a direct-API integration — cheap is the smallest/fastest model, standard is the default work model, frontier is the highest-reasoning production model, apex is reserved for the single batched tie-break call.

Nothing about this sits between OpenCode and your model provider. It's a policy file OpenCode's skill tool loads at dispatch time — no proxy, no network hop, nothing to route around if you remove the skill directory.

Install for OpenCode

Free. MIT. Nothing installs, nothing proxies your traffic.

Option 1 — skills CLI

This is the same generic skills CLI referenced on the main site. We haven't independently verified which of OpenCode's recognized skill directories it targets — if it writes to .claude/skills/, that path is one of OpenCode's documented discovery locations and needs no change; if you're unsure, use Option 2 below and place it under .opencode/skills/ directly.

Option 2 — copy the skill directory directly

Clone or download undercutsh/firstpass, then copy skills/firstpass/ into .opencode/skills/firstpass/ at your project root (or ~/.config/opencode/skills/firstpass/ to make it available globally). OpenCode's skill discovery walks up from your working directory to the git worktree root, so it also works dropped anywhere in between.

Validate it yourself

1. 1Read the published results — testing/README.md and the raw JSON in testing/results/
2. 2Reproduce for free, no API key, no spend: node src/main.js --mock
3. 3Run live with whichever provider you've already configured OpenCode against — a single-vendor smoke test is pennies

Zero lock-in. It's a policy OpenCode's skill tool reads — delete the directory and you're back to status quo.

Six flags assign a tier. Three triggers are the only way up.

This is a named, versioned spec — not internal logic we could quietly change. When the rubric changes (a new flag, a different cap), the version number moves and the changelog says why.

Every unit of work gets scored against six flags, then assigned the cheapest tier that can plausibly handle it. It only moves up a tier when something objective proves it needs to — a failed check, a measured disagreement, an explicit uncertainty flag. It never moves up on a guess, and it never moves back down mid-task.

 unit of work
      |
  6-flag rubric
      |
  +---+----+
  | 0 flags|--> cheap
  | 1-2    |--> standard
  | 3+/own |--> frontier
  +---+----+
      | fail x2 / disagree / uncertain
      v
  escalate +1 tier
  (residue only)
      |
  still unresolved?
      v
  single batched
  apex tie-break

Why it holds up

Flags steer. Verification and escalation decide.

Even when the rubric flags are scored imperfectly, the escalator still routes work to the right tier almost every time — because a wrong flag only ever costs one extra cheap attempt, never a wrong answer or a big bill.

Honest limits

What this doesn't do.

Not a proxy, not a gateway, doesn't auto-flag in production (OpenCode scores the flags itself — measured robust to getting them wrong, see above), and doesn't promise a dollar figure for your workload. Full list, with the finance/procurement note and the paid-tier disclosure, is on the main page.