Kiro already reads steering files and Agent Skills. Give it one that stops overpaying.

How this maps onto Kiro

Two confirmed paths: steering files, and native Agent Skills.

Kiro's documented, always-on mechanism is steering: plain markdown files at .kiro/steering/*.md (workspace scope) or ~/.kiro/steering/*.md (global scope), with workspace steering taking priority when the two conflict — per Kiro's own steering docs. Each file opens with a required YAML front-matter block (it must be the very first content in the file, no blank line before it) setting an inclusion mode: always (the default — loaded into every interaction), fileMatch (loaded only when a matching file pattern is open), or manual (pulled in on demand via a #steering-file-name reference). Kiro also auto-reads a root AGENTS.md as always-on, with no inclusion modes to configure.

Kiro separately ships native Agent Skills support — the same open, portable SKILL.md standard this repo already targets. Per Kiro's own skills docs, a skill is a folder containing a SKILL.md whose front matter must declare a name (lowercase letters, numbers, hyphens, matching the folder name) and a description Kiro uses to decide when to load it automatically. Skills live at .kiro/skills/<name>/SKILL.md (workspace) or ~/.kiro/skills/<name>/SKILL.md (global) — note Kiro requires the subfolder; it won't recognize a SKILL.md dropped straight into the skills directory. Undercut's skill folder is already named firstpass with that exact front-matter shape, so it copies across with nothing rewritten — see Install below.

Undercut's tier names — cheap, standard, frontier, apex — are generic on purpose. Kiro can run against more than one underlying model, so models.md gives the rule of thumb rather than a fixed slug for Kiro specifically. Nothing about this sits between Kiro and its model backend — it's plain markdown Kiro reads into context, no proxy, no network hop, nothing to route around if you remove the block.

Install for Kiro

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

Option 1 — install as a native Agent Skill (recommended, confirmed path, no edits needed)

Clone or download undercutsh/firstpass, then copy the skill folder straight into Kiro's workspace skills directory — the front matter already matches what Kiro's Agent Skills docs require, so nothing needs rewriting.

Option 2 — install once, for every workspace

Same folder, global scope. Kiro also reads ~/.kiro/skills/<name>/SKILL.md — useful if you want the policy available in every project without touching individual repos.

Option 3 — append as an always-on steering file

Prefer Kiro's steering mechanism instead? Save the policy under .kiro/steering/ with an inclusion: always front-matter block as the very first line of the file (Kiro requires no content before it), and it loads into every session the same way a project rules file would.

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. 3Open Kiro against a repo with the skill installed and check whether it references the routing policy before starting mechanical work

Zero lock-in. It's markdown text Kiro reads — delete the folder or the block 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 (Kiro scores the flags itself — measured robust to getting them wrong, see above), and doesn't promise a dollar figure for your workload. It's a routing policy, not a proof of savings on your workload — the benchmark numbers above show what the rubric measured on public tasks, not a guarantee for yours. Full list, with the finance/procurement note and the paid-tier disclosure, is on the main page.