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/luojiahai/code-by-wire/releasenpx skills add luojiahai/code-by-wire --skill releasegit clone --depth 1 https://github.com/luojiahai/code-by-wireWhat 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 | $0.00082 | $0.04953 |
| Opus 5 | $0.00041 | $0.02476 |
| Sonnet 5 | $0.00016 | $0.00991 |
| Haiku 4.5 | $0.00008 | $0.00495 |
Grade A, and why
release 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.
How it starts
The opening of the file, as written. The whole thing — 353 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Release
The maintainer drives releases in two phases, keyed off repo state. With the automation checkpoint below, one run can carry straight through both without stopping in between — pick the starting phase from what they asked and the repo state, then keep going as far as they authorize.
- Phase 1 — "bump version": the bump/changelog/PR are not on
mainyet. - Phase 2 — "release it": the bump PR has merged to
main; now tag and ship.
package.json version is the source of truth; the tag is v + that exact
string. CI internals, the botched-release runbook, and platform/signing notes
live at the end of this skill.
Orient first
Two things to settle before touching anything, especially on a bare /release
with no phase named.
Which phase? Compare the working version to the latest tag:
node -p "require('./package.json').version" # e.g. 0.1.5
git tag --sort=-v:refname | head -1 # e.g. v0.1.4
Version equals the latest tag → the last release shipped and nothing is
pending → Phase 1 (cut the next version). Version is ahead of the latest
tag → a bump already merged but isn't tagged → Phase 2 (tag and ship). When
unsure, git log --oneline --first-parent vLAST..HEAD shows what's unreleased.
gh defaults to the wrong host. This repo lives on personal GitHub, but gh
defaults to the work host, so every gh call needs both the host and the repo:
GH_HOST=github.com gh <cmd> -R luojiahai/code-by-wire ...
Plain git push/git tag are fine — only gh needs the prefix.
How far to take it
Merging, tagging, and publishing are each hard to reverse or visible to others, so they still deserve a deliberate go-ahead — but asking for that go-ahead three separate times, in three separate messages ("merge", then later "release it", then later "publish"), is pure friction when the maintainer already knows how far they want this to go. Collapse it: right after orienting, before touching anything, ask once via AskUserQuestion how far to drive this release. Scope the options to what's actually left from the phase you're starting in:
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 · 353 lines · 82 tokens per session scan A c5192d9c22cb
release is a skill published in the GitHub repository luojiahai/code-by-wire (74 stars, last pushed 7d ago), licensed MIT. It adds 82 tokens to every session and 4,953 once invoked, about $0.0004 per session on Opus 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 2026-08-30.
Other skills, from other repositories
bump-version
Use when bumping the Wayland version: update package.json, run checks, branch, commit, push, create PR, wait for merge, tag release.
vibetime-release
Use for VibeTime release work, including bumping the date-based version, creating and pushing release tags, triggering GitHub Actions CI packaging, creating or updating GitHub Releases with notes from the previous version, and verifying uploaded macOS/Windows assets. Trigger on requests such as 发版, 发布新版, release, bump…
chrome-release-verify
End-to-end Chrome security backport for an Electron release branch. Given a Chrome Releases blog URL and a branch (e.g. 41-x-y), determines which CVE fixes are missing from the actual synced source, writes the cherry-pick patches locally, validates them with e sync --3 + lint --patches, then pushes a single PR. Use…
changelog-draft
Generate a reviewable changelog draft from PRs merged in a release range. Extracts explicit CHANGELOG markers, classifies unmarked PRs, adds external contributor attribution, and outputs markdown + JSON artifacts. Does NOT mutate channelversions.json.
classify-changelog-pr
Reference guidance for classifying whether an unmarked PR should appear in the changelog and under which category. Used inline by the changelog-draft skill — not dispatched as a separate agent.
promote-feature
Promote a feature-flagged feature to Dogfood, Preview, or Stable in the Warp codebase. Use when a feature behind a FeatureFlag is ready to roll out to a broader audience, including wiring up the compile-time/runtime bridge and deferring flag cleanup safely.