Borrowing it
Nothing to install: this file belongs to drakulavich/kesha-voice-kit. Take a copy, put it at the same path in your own repository, and replace the rules that are about this project with yours.
curl -O https://raw.githubusercontent.com/drakulavich/kesha-voice-kit/main/.claude/skills/release-mechanics/SKILL.mdgit clone --depth 1 https://github.com/drakulavich/kesha-voice-kitWrote 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/drakulavich/kesha-voice-kit/release-mechanics)<a href="https://agentmods.dev/skills/drakulavich/kesha-voice-kit/release-mechanics"><img src="https://agentmods.dev/badge/skills/drakulavich/kesha-voice-kit/release-mechanics/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/drakulavich/kesha-voice-kit/release-mechanics"><img src="https://agentmods.dev/badge/skills/drakulavich/kesha-voice-kit/release-mechanics.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 | $0.00115 | $0.06759 |
| Opus 5 | $0.00057 | $0.03379 |
| Sonnet 5 | $0.00023 | $0.01352 |
| Haiku 4.5 | $0.00012 | $0.00676 |
Grade C, and why
release-mechanics scanned grade C with 2 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 2d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Recursive force deletehighDestructive command
rm -rf with a variable or a broad path is one typo away from removing the wrong tree.
SMOKE=/tmp/kesha-vX.Y.Z-smoke && rm -rf "$SMOKE" && mkdir "$SMOKE" && cd "$SMOKE" Makes network callslowCapability
Not a fault in itself. Listed so you know the mod talks to something, and to what.
5. Validate draft assets before un-drafting. Authenticated `gh release download` works on drafts; anonymous `curl` / `kesha install` 404s. Release drafts must include `SHA256SUMS`, `kesha-release-manifest.json`, one `*.s How it starts
The opening of the file, as written. The whole thing — 238 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Release mechanics
RELEASE PROCESS — CLI AND ENGINE ARE VERSIONED INDEPENDENTLY
package.json#version (CLI) and package.json#keshaEngine.version (engine, mirrored in rust/Cargo.toml) are decoupled. src/engine-install.ts downloads v${keshaEngine.version} with fallback to package.json#version.
Version drift gate: bun .github/scripts/check-versions.ts (bun run check:versions, CI "🔢 Check version drift") enforces:
keshaEngine.version === rust/Cargo.toml#version— one engine version stored twice; drift makeskesha installfetch the wrong source/release.package.json#version >= keshaEngine.version— CLI may lead for CLI-only patches, never lag.
CLI-only patch (docs, TS, plugin): bump package.json#version and the two server.json versions with it (rule 4 rejects the commit otherwise); leave keshaEngine.version + rust/Cargo.toml; PR CI uses the existing engine; merge; create a marker release:
CLI-only is allowed only when the changed CLI surface works against the already-published engine pinned by package.json#keshaEngine.version. If a CLI command delegates to a new engine subcommand, capability flag, feature behavior, or output contract, it is an engine release: bump package.json#keshaEngine.version, rust/Cargo.toml, and rust/Cargo.lock together. Before cutting any v*-cli marker, smoke-test new/changed CLI commands against the published pinned engine, not only a repo-local engine build. The v1.18.2-cli / v1.18.3-cli mistake was exposing kesha record while the pinned published engine was still v1.18.0 and did not implement kesha-engine record.
git tag -a vX.Y.Z-cli --cleanup=verbatim -F notes.md && git push origin vX.Y.Z-cli
npm view @drakulavich/kesha-voice-kit version # a few minutes, expect X.Y.Z
v*-cli is excluded from build-engine.yml and picked up by 🚀 Release (CLI) (release-cli.yml), which builds the Linux .deb/.rpm, creates the marker release as a draft with them plus SHA256SUMS, un-drafts it, then dispatches 📦 npm Publish and waits for it. Its body is the tag annotation followed by one generated line whose engine half is computed against the previous stable -cli tag's pin — the marker is how a new engine reaches users, and the line used to assert "(unchanged)" on exactly those releases (#788). A lightweight tag is still accepted; it just contributes no notes. Un-drafting from a workflow fires no release: published (a GITHUB_TOKEN event does not cascade), so the dispatch is explicit — which is also why the tag must be pushed by a human: a token-pushed tag triggers nothing. That is by design for the alpha lane, whose -cli tags have no GitHub release.
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 Changed e39e54b5cb96
- 8d ago Changed dc0d297c4f3a
- 12d ago First seen · 238 lines · 115 tokens per session scan C 4c86bf01add9
release-mechanics is a skill published in the GitHub repository drakulavich/kesha-voice-kit (73 stars, last pushed 2d ago), licensed MIT. It adds 115 tokens to every session and 6,759 once invoked, about $0.0006 per session on Opus 5. A static security scan graded it C with 2 findings (recursive force delete, 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
ship
Ship workflow: detect + merge base branch, run tests, review diff, bump VERSION, update CHANGELOG, commit, push, create PR. Use when asked to "ship", "deploy", "push to main", "create a PR", "merge and push", or "get it deployed". Proactively invoke this skill (do NOT push/PR directly) when the user says code is…
benchmark
Run performance + quality benchmarks. ASR reports WER, RTF, process memory, and throughput across engines/variants. Arguments include asr, tts, vad, diarize, asr-quick.
review-pr
Review a pull request for conceptual fit, architecture impact, adversarial failure modes, security risk, docs impact, regression risk, test coverage, and merge readiness. Use when asked to review a PR, check whether a PR is safe to merge, decide if more tests are needed, perform adversarial or security review, or…
commit-pr
Mandatory Codex/Copilot publication adapter for opencode-swarm. Use for every GitHub issue assignment that results in code changes, commits, pushes, draft PRs, PR body edits, PR readying, release notes, or CI closeout. Must be loaded before git push, gh pr create, gh pr edit, or gh pr ready. Routes to the single…
contributing
Codex adapter for the opencode-swarm contribution checklist. Use when preparing user-visible changes, release note fragments, local validation, branch hygiene, PR checklist items, or contribution-policy compliance before publishing work.
release-aur
Use when creating, updating, reviewing, or publishing Arch User Repository (AUR) packages. Runs /release-aur plan first, reviews PKGBUILD/.SRCINFO/build/namcap evidence, and only recommends /release-aur publish after a GO decision.