tailrocks/tailrocks-rust-skills
v0.28.1Apache-2.0
Strict Rust engineering workflows covering Axum, GraphQL, gRPC, TUI design, project baselines, review, refactoring, and remediation.
Changelog
0.28.1 - 2026-10-08
Applied the common active-package structure on branch
standardize/package-rewrite:
- Rewrote
plugin.jsonas the portable Agent Plugins 1.0.0 manifest. It is now the source of truth for name, version, and description. - Trimmed
.claude-plugin/plugin.jsonto name, version, and description. - Rewrote
.kimi-plugin/plugin.jsonwithskillsset to./skills/and a four-field interface block. - Removed the component marketplace file
.claude-plugin/marketplace.json. The centraltailrocksmarketplace is now the only catalog. - Removed the legacy host manifest
.codex-plugin/. Codex uses the portable manifest. - Removed the dead root file
catalog.json. Nothing referenced it. - Removed the root scripts runtime and the generated skill-definition
copies under
docs/. - Added
.alint.yml, pinned to the shared active profile. - Restructured
README.mdinto the eight required sections. - Replaced the generated docs index with the six standard guides
under
docs/. - Added
AGENTS.mdand.github/PULL_REQUEST_TEMPLATE.md. - Rewrote all 15
SKILL.mdfiles with the common body order in strict STE prose. - Corrected the binary-test and workspace-inheritance claims.
Labeled the
mod.rsand test-file rules as Tailrocks rules. - Split advisory duties: cargo-deny owns licenses, bans, and sources. Cargo-audit owns advisories.
- Removed the TypeScript crate-version helper. Version resolution uses native registry queries.
- Labeled Juniper, TanStack, Buf, and private gRPC use as Tailrocks choices.
- Regenerated CI with Velnor Actions 0.1.4.
- Replaced the
.github/CLAUDE.mdsymlink with a regular pointer file. Installers that reject symlinks now accept the package.
0.28.0 - 2026-10-07
Fifteen-skill package at commit 0317f100714dc66285c01c24f7967134375b5ac5
("adopt velnor-actions 0.1.0"). Five skills are model-selectable.
Ten skills are user-only and need an explicit human command.