qa

A quality-check procedure for the Radar project that chooses checks based on what changed. Type-checking checks whether the code's types fit together, while tests check expected behavior; visual tests check how changed screens or other interface elements actually look and work.

In plain words
What is it for?
Use it after code or interface changes to decide whether to run type-checking, unit tests, a build, and visual tests, then report the visual-test status.
Why use it?
It prevents unnecessary checks while still covering problems that code checks alone may miss, such as a screen that builds but renders incorrectly.

Command for Claude Code

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/skyhook-io/radar/qa
Clone the repo
git clone --depth 1 https://github.com/skyhook-io/radar

Made for: Claude Code.

Per session 18 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 550 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 $0.00018 $0.00550
Opus 5 $0.00009 $0.00275
Sonnet 5 $0.00004 $0.00110
Haiku 4.5 $0.00002 $0.00055

Measured 2d ago against content hash 582899de313b, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

qa 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 2d 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.

.claude/commands/qa.md · 38 lines

What it actually says

QA — radar

What "verify this change" means in radar. Run the checks that fit the delta; skip the rest. Don't run everything reflexively.

Code changed → type-check + tests

  • make tsc — frontend type-check. Never npm run lint (broken).
  • make test — unit tests.
  • make build — only before a PR, or when Go changed (does FE build + embed + binary; heavier, so not every round).

UI changed → /visual-test (lean toward it for feature-scale UI; skip is OK for small)

For a change that adds or meaningfully changes rendered surfaces — a new renderer / drawer / dialog / page / flow — lean toward running /visual-test: that's where it earns its keep, catching what tsc/test can't (does it actually render, read clearly, and behave). For small or non-visual changes (logic, utils, server/Go, types, refactors, copy, a tiny tweak) skipping is fine — don't ceremony-ize it. When unsure on a broad change, ask. Sometimes skipping a genuine UI change is still the right call — just make it a decision, and report it (below). (Aligns with radar's CLAUDE.md "consider visual-test" note.)

Always report visual-test status explicitly — never leave it ambiguous (a bare tsc ✓ · test ✓ with no visual-test entry reads as "did it run or not?"):

  • Ran → state the count and the screenshot dir as an absolute path (or a file:// URL) so the terminal linkifies it and the user clicks straight to the captures — e.g. visual-test: ran · 6 shots · /Users/.../radar/.playwright-mcp/visual-test/<run>/. Resolve $SCREENSHOT_DIR (from visual-test-start.sh) or $(pwd)/.playwright-mcp/... to a full path. Never a relative .playwright-mcp/... or a ~/... path — those don't reliably linkify in the terminal.
  • Skipped → say so with the reason: visual-test: skipped (no UI delta).

Pure backend / util / type / docs

make tsc / make test as relevant — no visual-test.

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. 2d ago First seen · 38 lines · 18 tokens per session scan A 582899de313b

Subscribe to this mod's changes

qa is a command published in the GitHub repository skyhook-io/radar (3,204 stars, last pushed today), licensed Apache-2.0. It adds 18 tokens to every session and 550 once invoked, about $0.0001 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 2026-08-30.