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 agentmods add skills/cloudflare/computer/debugging-computerd-fusenpx skills add cloudflare/computer --skill debugging-computerd-fusegit clone --depth 1 https://github.com/cloudflare/computerWhat 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 | $0.00120 | $0.03915 |
| Opus 5 | $0.00060 | $0.01958 |
| Sonnet 5 | $0.00024 | $0.00783 |
| Haiku 4.5 | $0.00012 | $0.00392 |
Grade A, and why
debugging-computerd-fuse scanned grade A with 1 finding 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.
Makes network callslowCapability
Not a fault in itself. Listed so you know the mod talks to something, and to what.
curl -sf http://127.0.0.1:18080/health How it starts
The opening of the file, as written. The whole thing — 376 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Debugging computerd against real FUSE
This skill captures the recipe for isolating bugs in computerd that only
fire when the kernel-FUSE backend is the one wired up. Most unit
tests run against SQLiteTestStorage and an in-process VFS or the
userspace FUSE_MOUNT=shim polling driver — neither exercises the
kernel-FUSE callback path. Several production-only bugs (write
buffer never spilled to the VFS, spawn(cwd=/mount) deadlocks)
have shipped to next and not been caught by the regular suites
because the harness is hard to drive end-to-end from a dev box.
The workflow here bypasses workerd, wrangler, and the
vitest-pool-workers harness. It boots the computerd SEA binary in a real
docker container with /dev/fuse and SYS_ADMIN, and drives its
capnweb endpoints from a plain Node script.
When to reach for this
- A user report of "wrote a file via FUSE, RPC reads see 0 bytes."
- A user report of "exec hangs forever" or "computerd
/healthstops responding after the first shell call." - Adding a new computerd op (FUSE callback, sync RPC, runner feature) and you want a smoke test that doesn't depend on Docker-in-DO or Cloudflare Containers.
- The harness vitest suite (
packages/computer/src/test-harness/) hangs atws.shell.execagainst the docker container and you need to attribute the wedge to computerd vs workerd vs the network shim. - Confirming a fix actually settles in the deployed binary: the SEA build is what wrangler ships, and reproducing locally with the same binary is the only way to be sure.
What you need
- Docker, with
/dev/fuseand--privilegedavailable. In Pi's sandbox this works through DinD; on a Linux dev box it works natively; on macOS you needcolima start --vm-type=vz(the default qemu vm doesn't surface/dev/fuse). - A fresh computerd binary at
artifacts/computerd/computerd-linux-x64. Build it withnpm run build:bin --workspace @cloudflare/computerd, or justnpm run build:allfrom the repo root if you need the docker image too.
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.
- 2d ago First seen · 376 lines · 120 tokens per session scan A 2e64f7972528
debugging-computerd-fuse is a skill published in the GitHub repository cloudflare/computer (8,834 stars, last pushed 4d ago), licensed MIT. It adds 120 tokens to every session and 3,915 once invoked, about $0.0006 per session on Opus 5. A static security scan graded it A with 1 finding (makes network calls). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.
Other skills, from other repositories
systematic-debugging
Use when encountering any bug, test failure, or unexpected behavior, before proposing fixes.
next-cache-components-adoption
Turn on Cache Components in a Next.js app and resolve the blocking routes it surfaces. Use when the user wants to enable, adopt, or migrate to Cache Components, flip the cacheComponents flag, work through a flood of blocking-prerender / instant validation errors, run the cache-components-instant-false codemod, or…
babysit-pr
Babysit a GitHub pull request after creation by continuously polling review comments, CI checks/workflow runs, and mergeability state until the PR is merged/closed or user help is required. Diagnose failures, retry likely flaky failures up to 3 times, auto-fix/push branch-related issues when appropriate, and keep…
imagegen
Generate or edit raster images when the task benefits from AI-created bitmap visuals such as photos, illustrations, textures, sprites, mockups, or transparent-background cutouts. Use when Codex should create a brand-new image, transform an existing image, or derive visual variants from references, and the output…
cpu-profile-analysis
Analyze V8/Chrome CPU profiles (.cpuprofile) and DevTools trace files (Trace-.json). Use when: profiling performance, investigating slow functions, comparing code paths, finding bottlenecks, analyzing timeToRequest, understanding call trees from sampling profiler data, analyzing layout/paint/rendering, investigating…
next-cache-components-optimizer
Drive a Next.js route to instant navigation by setting up an agentic loop, under Cache Components / PPR, on initial load (hard navigation) and client-side navigation (soft navigation). Encode the goal as a failing @next/playwright instant() e2e and work it to green, one verified route at a time; the shipped test then…