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 vincentkoc/dotskills --skill openclaw-bun-upgradegit clone --depth 1 https://github.com/vincentkoc/dotskillsWrote 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/vincentkoc/dotskills/openclaw-bun-upgrade)<a href="https://agentmods.dev/skills/vincentkoc/dotskills/openclaw-bun-upgrade"><img src="https://agentmods.dev/badge/skills/vincentkoc/dotskills/openclaw-bun-upgrade/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/vincentkoc/dotskills/openclaw-bun-upgrade"><img src="https://agentmods.dev/badge/skills/vincentkoc/dotskills/openclaw-bun-upgrade.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 | $0.00040 | $0.01290 |
| Opus 5.5 | $0.00016 | $0.00516 |
| Sonnet 5.5 | $0.00008 | $0.00258 |
| Haiku 4.5 | $0.00004 | $0.00129 |
Grade A, and why
openclaw-bun-upgrade 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 21d 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 — 119 lines — stays where its author put it; the contents beside it link to each section on GitHub.
OpenClaw Bun Upgrade
Purpose
Move an authorized Linux OpenClaw service to one verified Bun artifact while preserving its deployment owner, application source, state and active work. Keep source synchronization, build publication and runtime promotion as separate outcomes.
When to use
- Updating an OpenClaw Bun fork or selecting a daily build for a service.
- Building and qualifying a pinned Bun artifact for staged host promotion.
- Auditing how a service selects Bun, or preparing its runtime rollback.
The build and service recipes below cover Linux ELF and systemd deployments. For macOS or Windows, report the missing platform-specific qualification and service-owner route; do not apply Linux commands to those hosts. Source ancestry and immutable artifact selection principles still apply.
An audit does not authorize deployment. A runtime upgrade does not authorize an OpenClaw update, state migration, service-owner migration or automatic scheduler.
Workflow
- Bind scope and coordination. Name the hosts, deployment order, exact source, requested actions and completion conditions. Read each host's current owner instructions. If another recovery thread owns the host, exchange active command, lease and reservation status; wait for its explicit release before host commands or relay use. A release resolves coordination, not admission.
- Audit actual selection. Resolve service UID/GID and NSS home, native versus
external owner, unit plus drop-ins, effective executable and live PID/start
identity, application source/build, config revision and schema. Inspect native
runtime pins, update runs and rollback custody. Shell
bun --version, PATH, an oldcurrentlink or a campaign label is not running-service proof. - Freeze source and artifact. Read Build qualification. Reuse qualified objects and one artifact. Preserve upstream ancestry with a normal merge when requested; verify reachability rather than judging the PR's displayed diff. Freeze the full commit/tree before building. A newer daily candidate does not invalidate work on an already selected commit.
- Qualify the candidate. Match CPU baseline and ELF requirements to every target. Test the changed Bun bug class and actual OpenClaw consumers, with candidate identity inside workers. Scope carried proof to unchanged inputs. A successful build or generic smoke test alone is insufficient.
- Prepare promotion. Read Native promotion. Stage a versioned artifact, verify hash/ownership as the service account and establish rollback through the same deployment owner. Check the directory that must permit atomic replacement, not only the selected file. Do all expensive build and preflight work before acquiring a short maintenance lease.
- Admit and activate once. Use the deployed owner's native preserving admission, operation lock and service lifecycle. Live work blocks promotion unless that owner's exact durable continuation contract proves otherwise. Do not generalize a previously safe paused task or reuse another host's waiver. Lease, process or configuration drift stops activation. Uncertain mutation means reconcile the existing attempt; do not replay it.
- Verify the successor. Prove its actual executable hash, UID/cgroup and new process identity; unchanged application/config/schema when only Bun changed; restored admission; preserved task generations, payloads and bindings; health and an affected real consumer. Inspect relevant new-process errors. Report synthetic proof separately from live delivery or a naturally observed GC cycle.
- Promote the next host only after the first passes. Each host gets fresh ownership and admission proof; artifact compatibility may be reused. If blocked, report the precise failed gate and release command/lease/reservation custody. Preserve another owner's recovery material. Do not keep polling after a bounded check or retry an unchanged failure.
- Close out. Report source sync, artifact qualification, per-host deployed identity, remaining blockers and checkout disposition separately. Release the task's builder and child processes. Keep the previous runtime while rollback needs it, and the candidate while another authorized rollout remains unfinished. Dispose of task scratch through its owner when no unfinished work needs it.
What ships with it
4 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.
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.
- 21d ago First seen · 119 lines · 40 tokens per session scan A 2aa50f8fa52d
openclaw-bun-upgrade is a skill published in the GitHub repository vincentkoc/dotskills (109 stars, last pushed yesterday), licensed MIT. It adds 40 tokens to every session and 1,290 once invoked, about $0.0002 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 2026-09-26.
Other skills, from other repositories
gemigo-cli
Use when the user wants to publish an already-built static website or front-end app through GemiGo, such as a Vite/React/Vue static build, plain HTML/CSS/JS page, landing page, demo, docs site, or small browser app, and get a hosted public URL.
architecture-diagram
Generate dark-themed SVG diagrams of software systems and cloud infrastructure as standalone HTML files with inline SVG graphics. Semantic component colors (cyan=frontend, emerald=backend, violet=database, amber=cloud/AWS, rose=security, orange=message bus), JetBrains Mono font, grid background. Best suited for…
serving-llms-vllm
Serves LLMs with high throughput using vLLM's PagedAttention and continuous batching. Use when deploying production LLM APIs, optimizing inference latency/throughput, or serving models with limited GPU memory. Supports OpenAI-compatible endpoints, quantization (GPTQ/AWQ/FP8), and tensor parallelism.
tensorrt-llm
High-throughput LLM inference on NVIDIA GPUs.
stripe-projects
Provision SaaS services + sync creds via Stripe Projects.
lambdasim
Run the studio's per-project NLP Lambda path locally with lambdasim, haven's AWS Lambda stand-in: functions, invokes and response streams answered on this stack's nlpgo. Use when someone says 'haven lambda', 'lambdasim', 'haven up +lambda', 'haven sim lambda', 'NLP Lambda locally', 'per-project Lambda'…