evaluate

A no-write trial that replays recent merged changes, or a chosen before-and-after commit range, through the same checks used by hooks and continuous integration.

In plain words
What is it for?
Use it to assess recent changes in disposable worktrees, measure check impact, and inspect results as JSON or an optional output file.
Why use it?
It shows what the checks would have blocked and what setup cost to expect before changing the repository.

Command

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/vyuh-labs/dxkit/evaluate
Clone the repo
git clone --depth 1 https://github.com/vyuh-labs/dxkit
Per session 0 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 863 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 $0.00000 $0.00863
Opus 5 $0.00000 $0.00432
Sonnet 5 $0.00000 $0.00173
Haiku 4.5 $0.00000 $0.00086

Measured 2d ago against content hash ab25f10d22e5, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

evaluate 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.

docs/commands/evaluate.md · 66 lines

How it starts

The opening of the file, as written. The whole thing — 66 lines — stays where its author put it; the contents beside it link to each section on GitHub.

vyuh-dxkit evaluate

Run as: vyuh-dxkit <cmd> after npm install -g @vyuhlabs/dxkit, or npx @vyuhlabs/dxkit <cmd> for one-shot use. Examples on this page use the short form.

A zero-write trial. evaluate replays your repo's recent landings (the last few merged changes, or an explicit before/after ref pair) through the same deterministic guardrail gate the hooks and CI would run, and reports what the gate would have blocked plus what enabling dxkit costs (measured gate latency, interruption rate, setup). It answers "would dxkit help here, and what does it cost?" before you commit to installing anything.

evaluate is an advisory replay, not a gate. It runs the trial in disposable worktrees and writes nothing to your repository unless you pass --output. Its exit code is always 0 on a completed trial, even when a replayed landing would have been blocked. Scripts that want a pass/fail signal read the JSON (totals.blocked) or run guardrail check.

Usage

vyuh-dxkit evaluate [path] [--last-prs <n>] [--base <ref> --head <ref>]
                    [--preset security-only|full-debt] [--json]
                    [--redact] [--output <file>]

With no flags it replays a default window of recent landings and prints a terminal summary. Each landing takes roughly 30 seconds to a minute to replay, so a small window returns in a couple of minutes.

Flags

Flag Effect
--last-prs <n> Replay the last n merged changes (a positive integer)
--base <ref> Replay a single explicit before/after pair: the "before" ref
--head <ref> The "after" ref of the explicit pair (paired with --base)
--preset <name> Which findings block: security-only (default) or full-debt. Same presets as the loop Stop-gate
--json Emit the versioned evidence document to stdout, for tooling
--redact Strip finding evidence (locations, captured snippets) from the output, for sharing
--untrusted Treat the scanned source as attacker-controlled, matching a hosted PR gate's read-only scanner posture
--no-incremental Force a full scan on each replay instead of the diff-scoped default
--output <file> Write the evidence JSON to a file (an explicit, user-requested write; the repo itself receives nothing)
--verbose Print per-tool timing to stderr
path Repository to evaluate (defaults to the current directory)

Read the full file on GitHub · 66 lines

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 · 66 lines · 0 tokens per session scan A ab25f10d22e5

Subscribe to this mod's changes

evaluate is a command published in the GitHub repository vyuh-labs/dxkit (10 stars, last pushed 4d ago), licensed MIT. It costs nothing until one of its globs matches a file; then it loads 863 tokens. 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 2026-08-31.