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 simiancraft/simiancraft-skills --skill walk-the-floorgit clone --depth 1 https://github.com/simiancraft/simiancraft-skillsWrote 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/simiancraft/simiancraft-skills/walk-the-floor)<a href="https://agentmods.dev/skills/simiancraft/simiancraft-skills/walk-the-floor"><img src="https://agentmods.dev/badge/skills/simiancraft/simiancraft-skills/walk-the-floor/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/simiancraft/simiancraft-skills/walk-the-floor"><img src="https://agentmods.dev/badge/skills/simiancraft/simiancraft-skills/walk-the-floor.svg" alt="Reviewed on agentmods" width="80" height="20"></a>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 | $0.00218 | $0.01433 |
| Opus 5.5 | $0.00087 | $0.00573 |
| Sonnet 5.5 | $0.00044 | $0.00287 |
| Haiku 4.5 | $0.00022 | $0.00143 |
Grade A, and why
walk-the-floor 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 5d 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 — 94 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Walk the Floor
A list checker and fixer for a running environment. It wakes, reads a list of things that should
now be true, goes and looks at each one to the standard would a user notice, writes what it
found to a ledger, and when the environment is wrong it files an incident and hands the incident to
fix-github-issue. It knows nothing about who wrote the list or why.
Three things make it composable rather than a monitor:
- The floor is a directory.
list.jsonlin,ledger.jsonlout, both append only. A person, a loop, another skill, or the walker's own forge producer (it reads the code host, GitHub here, for merged pull requests) can all put items on the floor; the contract is inreferences/the-floor.md. - Two callback slots.
on-passandon-fail, each an executable the driver runs or a prompt the agent reads. Whoever starts the walker decides what a failure means to them; the walker does not. The issue burndown useson-failto stop its own merge queue. - The environment is config. Kind, base URL, health paths, how to read the deployed revision,
how to log in, which endpoints are safe to post to, and the standing sanity walks all live in
walk-the-floor.config.tsat the repository root. The skill never contains a project.
The standard
Sanity, not integration. Go to the thing and look. If the change is a label, read it; if it is behind a surface, do the ordinary user action that runs through it and see that it works; if there is no surface, hit the endpoint or read the database and say that is what you did. A site that boots, renders, and returns nothing for every search has no data, and only searching tells you. The walk proves a user would not hit a wall; it does not prove the change is correct.
Run it
From inside the target repository:
bun run <this-skill-dir>/walk.ts --dir <floor> --liveness-only --once # probe the base URL and stop
bun run <this-skill-dir>/walk.ts --dir <floor> --once # one full wake
bun run <this-skill-dir>/walk.ts --dir <floor> --every 10 # forever, every ten minutes
bun run <this-skill-dir>/walk.ts --dir <floor> --once --dry-run # probe and classify; no agent, no checkout, no list or ledger write, no callback, nothing filed
bun run <this-skill-dir>/walk.ts --dir <floor> --once --walker claude:claude-opus-5
bun run <this-skill-dir>/walk.ts --dir <floor> --once --base-url https://preview.example.com # walk this URL instead of the configured one
bun run <this-skill-dir>/walk.ts --dir <floor> --once --max-points 3 # cap the incident fix handed to a worker for this run
What ships with it
16 files beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
- lib/callbacks.ts 816 B runs code
- lib/config.test.ts 456 B runs code
- lib/config.ts 9.0 KB runs code
- lib/floor.ts 5.8 KB runs code
- lib/forge.ts 2.7 KB runs code
- lib/incident.ts 9.9 KB runs code
- lib/liveness.ts 1.8 KB runs code
- lib/prompts.ts 726 B runs code
- lib/revision.ts 1.9 KB runs code
- prompts/diagnose.md 2.1 KB
- prompts/walk.md 3.7 KB
- references/adopting.md 9.2 KB
- references/operating.md 5.9 KB
- references/the-floor.md 5.8 KB
- walk.source.test.ts 3.0 KB runs code
- walk.ts 22 KB runs code
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.
- 5d ago First seen · 94 lines · 218 tokens per session scan A df75290c9df7
walk-the-floor is a skill published in the GitHub repository simiancraft/simiancraft-skills (8 stars, last pushed 6d ago), licensed MIT. It adds 218 tokens to every session and 1,433 once invoked, about $0.0009 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 2026-10-20.
Other skills, from other repositories
testing-and-quality
Use throughout the build and before claiming any phase or the project "done", to run the four mandatory quality gates and verify them with evidence. Covers the test pyramid, Playwright + Chrome DevTools browser verification, WCAG AA, and the commit/finish gate wiring.
demo-review
Quality-checks a rendered demo video before it ships — extracts frames, grades each section against its successCriteria, and finds spinners, empty states, clipped text, exposed secrets, audio/video desync, and pacing problems. Use after a draft render, before a final render, or when asked whether a demo video is good…
test-native-extension
Validate a third-party control repo across four automated layers plus one printed manual recipe. Layer 1 asserts native-source structure (Android getName() and iOS +moduleName to manifest nativeModule; @ReactMethod / RCTEXPORTMETHOD to methods; no @ReactModule) plus load/init readiness (ReactPackage public no-arg…
test-site
Tests a deployed, activated Power Pages site at runtime using browser-based navigation, page crawling, and API request verification via Playwright. Use when the user wants to test, verify, or smoke-test their deployed site.
qa
Browser-based QA verification. Launches a real browser, navigates the app, clicks buttons, fills forms, and tests user flows. Works as a standalone skill or as a phase end condition in campaigns. Requires Playwright (optional dependency, graceful skip if not installed).
live-preview
Mid-build visual verification loop. Takes screenshots of components during construction, not just after. Catches visual regressions and invisible features before they compound. Requires Playwright or similar screenshot tool.