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.
npx skills add vyuh-labs/dxkit --skill dxkit-evaluategit clone --depth 1 https://github.com/vyuh-labs/dxkitWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/skills/vyuh-labs/dxkit/dxkit-evaluate)<a href="https://agentmods.dev/skills/vyuh-labs/dxkit/dxkit-evaluate"><img src="https://agentmods.dev/badge/skills/vyuh-labs/dxkit/dxkit-evaluate/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/skills/vyuh-labs/dxkit/dxkit-evaluate"><img src="https://agentmods.dev/badge/skills/vyuh-labs/dxkit/dxkit-evaluate.svg" alt="Reviewed on agentmods" width="80" height="20"></a>- NVIDIA SkillSpector warn
SkillSpector: 1 finding, up to medium
These are SkillSpector’s own severities. On a checked sample its high-severity flags on skills were ~96% false positives — a documented command, a public API, a “never do X” rule — so we show them as a caution to read, not a verdict. Why →
- medium MCP Rug Pull · line 48 npx commands without a version suffix (e.g. @1.0.0) create a rug-pull risk if the upstream server is compromised and publishes a malicious update.Fix: Pin the version: npx @scope/[email protected]
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.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00090 | $0.00665 |
| Opus 5.5 | $0.00036 | $0.00266 |
| Sonnet 5.5 | $0.00018 | $0.00133 |
| Haiku 4.5 | $0.00009 | $0.00067 |
Grade A, and why
dxkit-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 1mo 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.
How it starts
The opening of the file, as written. The whole thing — 53 lines — stays where its author put it; the contents beside it link to each section on GitHub.
dxkit-evaluate
This skill answers the pre-adoption question honestly: what would dxkit's
gate have done on this repo's real history, and at what cost? It runs the
real guardrail — the same code path the installed Stop-gate and CI gate use —
against historical ref pairs, in disposable temp worktrees. Nothing is
written to the repo: no .dxkit/, no baseline, no hooks.
Running the trial
The default replays the last 10 landings (merged PRs / squash commits) of the current branch:
vyuh-dxkit evaluate # last 10 landings, security-only posture
vyuh-dxkit evaluate --last-prs 20 # more history
vyuh-dxkit evaluate --preset full-debt # also gate test gaps + quality regressions
vyuh-dxkit evaluate --base origin/main~5 --head origin/main # one explicit range
vyuh-dxkit evaluate --json # the versioned evidence document
vyuh-dxkit evaluate --json --redact # shareable: file paths + lines stripped
Reading the result
- Blocked landings list the exact net-new findings the gate would have returned for repair. Dependency findings on old landings carry an anachronism note (advisories are current-day) — say so when summarizing.
- A clean replay is a positive result, not an empty one: it means the repo's existing debt would be grandfathered and the gate would have stayed out of the way. Present it that way — the gate's value is blocking the regression that hasn't happened yet, and the trial just measured its false-block behavior on real history.
- The costs section is measured, not modeled: gate latency comes from this trial on this machine, the interruption rate from the repo's own history. Quote those numbers; do not invent others.
- Some finding classes are structurally excluded from a ref-vs-ref replay (duplication, test gaps, custom checks) — the output discloses this. Do not claim the trial covered them.
Next steps to offer
- Clean history and the user wants to see a block happen:
npx -y @vyuhlabs/dxkit@latest demo loop-guardrail(20 seconds, fixture repo, their repo untouched). - Ready to enable:
npm init @vyuhlabs/dxkit -- --claude-loop --yes, thenvyuh-dxkit baseline create. Everything is reversible viavyuh-dxkit uninstall.
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.
- 1mo ago First seen · 53 lines · 90 tokens per session scan A 0834208c3bac
dxkit-evaluate is a skill published in the GitHub repository vyuh-labs/dxkit (10 stars, last pushed 1mo ago), licensed MIT. It adds 90 tokens to every session and 665 once invoked, about $0.0004 per session on Opus 5.5. 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 2026-08-31.
Other skills, from other repositories
Code Coverage Analysis
Measure and enforce test coverage with Istanbul/nyc, c8, Jest, and Vitest. Covers branch versus line coverage, per-directory thresholds, CI gates, and correctly excluding generated code from reports.
release-habit-hooks
Cut a new release of the habit-hooks packages. Use when asked to release, publish, or bump the version. Reviews what lands, enforces the in-sync versioning rule, validates the changelog, and drives the tag-triggered PyPI publish.
habit-hooks-review
Spawn a reviewer sub-agent to assess a change set against habit-hooks's coding principles. Use AFTER habit-hooks reports clean — habit-hooks catches structural smells; this catches what it cannot (correctness, tests, design, missed edge cases).
duo-build
Build an issue or feature interactively toward an indexed shared goal, with automatic read-only peer review; resume the same run across Claude Code and Codex.
specflow
Spec-driven development with executable contracts.
specflow-loop-selector
Choose the lightest Specflow workflow that covers an ask’s size, uncertainty and risk, then select a loop only when needed. Use for work triage, proportionate verification, loop selection, or explicit spec-build, feature-build and Gate D requests. A lightweight decision does not start a loop.