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/yoheimuta/protolint/releasenpx skills add yoheimuta/protolint --skill releasegit clone --depth 1 https://github.com/yoheimuta/protolintWhat 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 | $0.00047 | $0.01063 |
| Opus 5 | $0.00023 | $0.00531 |
| Sonnet 5 | $0.00009 | $0.00213 |
| Haiku 4.5 | $0.00005 | $0.00106 |
Grade A, and why
release 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 -s https://pypi.org/pypi/protolint-bin/json | jq -r .info.version How it starts
The opening of the file, as written. The whole thing — 106 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Releasing protolint
One git tag fans out to six channels. release.sh only pushes the tag; everything
else is CI. The job is mostly verifying that each channel really published,
because they fail independently and silently.
1. Preflight
Run from a clean master that is in sync with origin:
go build ./... && go test ./...
(cd bdist/js && npm ci --ignore-scripts && npx tsc && npx eslint . && npx prettier --check .)
.github/workflows/go.yml and the JS workflows only run on pull_request, so
nothing has verified master as a whole. Do not skip this.
2. Pick the version
Every channel takes its version from the git tag — .goreleaser.yml ldflags,
build.gradle from GITHUB_REF_NAME, and npm version from-git. Nothing needs
bumping by hand; the tag is the version.
Look at what changed since the last tag, and remember the npm package ships to a different audience than the Go binary:
- patch — Go-side fixes, dependency bumps, docs
- minor — new lint rules, or anything that changes what npm consumers need:
a raised
engines.node, a module-system switch, renamed install scripts
3. Tag
./release.sh vX.Y.Z "Release vX.Y.Z"
4. Watch the pipeline
goreleaser runs on the tag push. publish_npm and publish_wheel then trigger
off its completion, not its success — so a green goreleaser does not mean the
downstream jobs passed.
gh run watch <goreleaser-run-id> --exit-status
gh run list --workflow publish_npm.yml --limit 1
gh run list --workflow publish_wheel.yml --limit 1
5. Verify every channel
Check the registries, not the workflow badges.
gh release view vX.Y.Z --json assets --jq '.assets|length' # expect 9
npm view protolint version
curl -s https://pypi.org/pypi/protolint-bin/json | jq -r .info.version
curl -s https://repo1.maven.org/maven2/io/github/yoheimuta/protoc-gen-protolint/maven-metadata.xml | grep release
Maven Central publishes automatically via the Central Publisher Portal step in
goreleaser.yml (publishing_type=automatic). No manual staging promotion.
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 · 106 lines · 47 tokens per session scan A ff53c54a95e8
release is a skill published in the GitHub repository yoheimuta/protolint (697 stars, last pushed 18d ago), licensed MIT. It adds 47 tokens to every session and 1,063 once invoked, about $0.0002 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
slopless
Use Slopless to review English Markdown for deterministic AI and human slop signals, including vague phrasing, formulaic prose, weak rhythm, filler, and cliches.
suiko
日本語文書のAI由来の均一さ、翻訳調、不自然さ、論旨、読解負荷を、決定的なRust CLIと目視で診断し、依頼に応じて書く・直す。Use when the user explicitly mentions suiko, asks whether Japanese text looks AI-generated, requests a naturalness score or final readability pass, wants a Japanese meeting note, report, guide, proposal, email, slide outline, blog, note, or essay made…
updater
Check Claude Code changelog for lintable changes and implement cclint updates. Use when a new Claude Code version ships, when CLAUDE.md claudecodelastupdated is stale, or for version gap detection, domain classification, and parallel agent dispatch across CUE schemas, hook events, known tools, and lint rules.
golang-grpc
Production gRPC in Go: protobuf layout, codegen, interceptors, deadlines, error codes, streaming, health checks, TLS, and testing with bufconn.
slop-chop
Remove AI writing tells from text with the slop-chop CLI. Use when the user wants to clean up a draft, strip em-dashes, semicolons, and stock buzzwords, enforce a spelling dialect, or make text read like a person wrote it instead of a chatbot. Also use before handing back a written draft, or to flag slop in Markdown…
stop-slop
Remove AI writing patterns from prose. Use when drafting, editing, or reviewing text to eliminate predictable AI tells.