A release workflow that decides whether changes are ready to publish and selects a version using semantic versioning. Semantic versioning uses major, minor, and patch numbers to describe the impact of changes.

Use it to review changes since the last release, choose the appropriate version bump, and follow the repository's commit, pull-request, tag, and release process.

It removes guesswork from deciding when to release and how much to change the version number.

cut-release — Skill for Claude Code and Codex by coeusyk: A release workflow that decides whether changes are ready…

Install

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.

agentmods

npx agentmods add skills/coeusyk/inference-x/cut-release

Any agent

npx skills add coeusyk/inference-x --skill cut-release

Clone the repo

git clone --depth 1 https://github.com/coeusyk/inference-x

Made for: Claude Code, Codex.

Wrote 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.

README.md

[![agentmods](https://agentmods.dev/badge/skills/coeusyk/inference-x/cut-release.svg)](https://agentmods.dev/skills/coeusyk/inference-x/cut-release)

Your own site

<a href="https://agentmods.dev/skills/coeusyk/inference-x/cut-release"><img src="https://agentmods.dev/badge/skills/coeusyk/inference-x/cut-release.svg" alt="Measured on agentmods" height="20"></a>

Per session 38 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.

When invoked 1,009 The whole file, excluding the scripts and references it only reads on demand.

Security scan A 0 findings. Scan, not verified.

Origin original No closer match found in the catalogue.

Token cost

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	$0.00038	$0.01009
Opus 5	$0.00019	$0.00504
Sonnet 5	$0.00008	$0.00202
Haiku 4.5	$0.00004	$0.00101

Measured 5d ago against content hash 9754c52113a2, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

cut-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 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.

.claude/skills/cut-release/SKILL.md · 80 lines

How it starts

The opening of the file, as written. The whole thing — 80 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Cut a release

Use when asked to release, version-bump, or judge whether accumulated develop work is "worth" a release.

1. Decide the version — semver.org rules

Given pyproject.toml's current version and the set of OpenSpec changes merged to develop since the last release tag:

MAJOR (X.0.0) — any backward-incompatible change to a public contract: a route removed/renamed, a response schema field removed or repurposed, a behavior change existing callers relied on.
MINOR (X.Y.0) — new backward-compatible functionality: a new route, a new optional field, a new capability that doesn't touch existing contracts.
PATCH (X.Y.Z) — backward-compatible bug fixes only, no new surface.

Bump once for the whole batch, not once per change — take the highest tier any single change in the batch requires. A batch containing one MINOR change and three PATCH changes is still MINOR.

This repo treats 0.x releases as real, numbered, user-facing releases (see existing tags) — don't withhold a release just because major version is still 0.

2. Confirm release-worthiness

Not every merge needs a release. Cut one when develop has diverged from the last release tag with at least one user-visible change (new endpoint, behavior fix, dependency-facing change). Pure internal refactors with zero external surface change (like an engine-driver swap with no API change) are still worth bundling into the next release rather than triggering one alone — check git log <last-tag>..develop --oneline for what's accumulated.

3. This repo's release pattern (mirror it, don't reinvent)

Verified from git history (git show --stat on prior release-prep commits, e.g. 5b81605, 51c764f):

Feature work already landed on develop via normal feat/→develop PRs (each with its own feat(...) commit + separate chore(openspec): archive <id> commit).
Branch chore/release-vX.Y.Z from develop.
Bump only pyproject.toml's version field and regenerate uv.lock (uv lock) — nothing else. CHANGELOG.md is NOT touched by this commit; there is no repo precedent for closing out ## [Unreleased] into a versioned section, and none should be invented without asking.
Commit as chore(release): prepare Phase A vX.Y.Z (adjust the phase name if the repo has moved past Phase A).
PR chore/release-vX.Y.Z → develop, wait for the required checks CI job, merge.
PR develop → main (title: Release vX.Y.Z: <short theme>), wait for checks, merge. main requires code-owner review per the "Protect Main" ruleset — if merge is blocked on review, stop and ask rather than overriding it.
On main, tag: git tag -a vX.Y.Z -m vX.Y.Z && git push origin vX.Y.Z.
gh release create vX.Y.Z --title "vX.Y.Z — <theme>" --notes-file ... with a hand-authored narrative body (NOT a copy of CHANGELOG.md's terse bullets) structured as:
- Title: InferenceX vX.Y.Z — <thematic summary>
- ### What changed — one subsection per OpenSpec change, each a prose paragraph naming the change id and explaining what changed and why, in the style of prior releases (gh release view v0.3.0 --json body for reference).
- Optional ### Also catch-all for minor/doc-only additions.
- **Full Changelog**: https://github.com/<owner>/<repo>/compare/vPREV...vNEW

Read the full file on GitHub · 80 lines

Changes

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 · 80 lines · 38 tokens per session scan A 9754c52113a2

Subscribe to this mod's changes

cut-release is a skill published in the GitHub repository coeusyk/inference-x (2 stars, last pushed 25d ago), licensed MIT. It adds 38 tokens to every session and 1,009 once invoked, about $0.0002 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 2026-08-31.

Git and workflow LLM Releases inference openai-compatible self-hosted sse

sglang-cherrypick

Trigger the bot-cherry-pick workflow for a batch of merged PRs onto a release branch and monitor each run to completion. Use when an SGLang release manager asks to cherry-pick a list of PRs to a release branch.

sgl-project/sglang · 56 tokens

groq-inference

Ultra-fast LLM inference on custom LPU hardware. OpenAI-compatible API at api.groq.com. Lowest latency in the industry (500-1000+ tok/s). Supports chat completions, vision, audio (Whisper STT + TTS), tool calling, JSON mode, and streaming. Free tier available. Inference only — no training.

synthetic-sciences/openscience · 77 tokens

pre-release-check

The final gate before a release is cut — go green, reconcile the release PR against what actually landed, sweep what CI can't see, check the roadmap and ADR statuses, optionally deep-review, then return a ship-or-hold verdict.

willdady/platypus · 51 tokens

fireworks-ai-inference

Fast inference and fine-tuning platform with serverless and on-demand GPU deployments. OpenAI-compatible API for chat completions, embeddings, function calling, vision, and structured output. Supports SFT, DPO, and RL fine-tuning. SOC2 + HIPAA compliant.

synthetic-sciences/openscience · 62 tokens

add-dashboard-orpc-procedure

Verify permission guards when creating, modifying, or reviewing any oRPC procedure file in packages/xinity-ai-dashboard/src/lib/server/orpc/procedures/.

xinity-ai/xinity-ai · 39 tokens

run-tests

End-to-end verification that every test suite in the workspace runs green locally. Use when validating dependency changes, refactors, or anything else that could plausibly affect runtime behavior across packages.

xinity-ai/xinity-ai · 40 tokens