mcp-repo-release-tools
Summary: This MCP server exposes repo-release-tools (rrt) capabilities for inspecting release policy, versioning, git workflow, changelog, validation, and health in a repository.
Read and inspect resolved
[tool.rrt]configuration and check whether hooks, workflows, and artifact protection are wired up.Inspect recorded snapshot locks (health, drift, tree, artifacts) and run live checks for release consistency, folders, docs, EOL, and upstream sync.
Get authoritative current versions and preview or apply version bumps across version groups, pins, changelog, lockfiles, and generated assets.
Validate branch names and Conventional Commit subjects before use, and create policy-compliant branches.
Read the Unreleased or full changelog.
Use git helpers for dirty-tree/generated-path detection, commits since a ref, landing branches, and publishing squashed snapshots.
Run repository-declared gate commands and judge them against expected exit codes.
Render human-facing UI dashboards/forms for health, versions, doctor results, tree, locks, and initialization.
Provides tools for Git repository release management, including branch creation, version bumping, changelog updates, and commit subject validation.
repo-release-tools
repo-release-tools keeps release policy boring in the best possible way.
Use it from GitHub Marketplace when you want CI to validate branch names, commit subjects, and changelog policy. Install it from PyPI when you want a local CLI, hook integration, version bumps, and release-branch automation.
GitHub Marketplace action: https://github.com/marketplace/actions/repo-release-tools-policy-checks
PyPI package: https://pypi.org/project/repo-release-tools/
Choose your entry point
Use the GitHub Action for CI policy checks
Choose the action if you want pull requests and pushes to fail fast when a repo drifts from your release policy.
validates branch names such as
feat/add-parservalidates Conventional Commit subjects
validates changelog policy in CI
optionally checks that the working tree stays clean
can run
rrt doctoras a pre-release health gate
- uses: actions/checkout@v6
with:
fetch-depth: 0
- uses: Anselmoo/repo-release-tools@v1.20.0
with:
check-branch-name: "true"
check-commit-subject: "true"
check-changelog: "true"See the full action guide: https://anselmoo.github.io/repo-release-tools/action/
See the full CLI and commands reference: https://github.com/Anselmoo/repo-release-tools/blob/main/docs/commands/rrt-cli.md
Use the Python package for local workflow automation
Choose the package if you want the developer-side tools: branch helpers, version bumps, config inspection, pre-commit hooks, and release automation. The Python package is published on PyPI and has a CI counterpart in the GitHub Action guide.
pip install repo-release-tools
rrt init
rrt branch new feat "add parser"
rrt git commit "add parser"
rrt git doctor
rrt bump patchOr run the CLI without installing it permanently:
uvx repo-release-tools branch new feat "add parser"If rrt is already installed and you want the bundled agent skill for Copilot,
Claude, or Codex, install it with:
rrt skill install --target copilot-local
rrt skill install --target claude-local --target codex-local
rrt skill install --target codex-global --dry-runFor basic versioning, bump and ci-version can run without [tool.rrt] by
auto-detecting repo-root pyproject.toml, package.json, Cargo.toml,
.rrt.toml, or .config/rrt.toml.
If multiple version files are found, they are updated together. Explicit config
is for the nice extras: grouped releases, changelog paths, release branches,
lock commands, generated files, and custom patterns.
Version targets also support common language/project files such as Python
(pep621, python_version), Node/JS/TS (package_json), Go (go_version),
Rust (cargo_toml), and .NET (csproj) so multi-language repositories can
keep their release versions aligned. A mcp_server_json target keeps an MCP
Registry server.json (top-level version, each package's version, and
any oci package's image tag) in sync alongside a project's primary target.
Related MCP server: mcp-creator
Changelog workflows
Pick the style that matches how your repository lands changes.
incremental (default) — for teams that maintain changelog entries during development.
rrt-update-unreleasedandrrt-changeloghooks stay active.The GitHub Action resolves
changelog-strategy: autotoper-commit.rrt bumpdefaults toauto.
squash — for repositories that squash many commits into one PR merge.
Changelog write and check hooks skip enforcement.
The GitHub Action resolves
changelog-strategy: autotorelease-only.rrt bumpdefaults togenerate.
Minimal config:
[tool.rrt]
release_branch = "release/v{version}"
changelog_file = "CHANGELOG.md"
changelog_workflow = "incremental" # or "squash"
[[tool.rrt.version_targets]]
path = "pyproject.toml"
kind = "pep621"Native config is also supported in package.json ("rrt": { ... }) and
Cargo.toml ([package.metadata.rrt] / [workspace.metadata.rrt]). Go repos
should use .rrt.toml or .config/rrt.toml.
What the project includes
rrtCLI for branches, bumps, config inspection, and Git helpersrrt-hooksforpre-commit,lefthook,husky, and CI validationa reusable GitHub Action in
action.ymlbundled agent skills for
uvxand installed-CLI workflowsdocs for branch policy, hook setup, and release workflows
Get your AI agent to actually use rrt
If you use Claude Code, Copilot, Cursor, or Codex on a repository that has rrt
configured, the agent will happily reimplement what rrt already does — hand-editing a
version string in three files, inventing a branch name the pre-commit hook then rejects,
or writing a changelog entry in the wrong section. Not because it lacks the tools, but
because nothing told it these are the tools for this job.
Two things fix that. Do both.
1. Tell your agent, once, in its instruction file
Paste the block from Agent instruction snippet below into
your repository's CLAUDE.md, AGENTS.md, or .github/copilot-instructions.md. This is
the highest-leverage step by a wide margin: it is read on every session, before the agent
has formed a plan, and it costs nothing at runtime.
2. Connect the MCP server (optional, but better for anything that writes)
uv add "repo-release-tools[mcp]"Then add .mcp.json at the repository root — Claude Code picks it up on next start:
{
"mcpServers": {
"rrt": { "type": "stdio", "command": "uv", "args": ["run", "rrt-mcp"] }
}
}With the server connected, most tools give the agent typed JSON instead of terminal
output it has to parse (the four lock readers and rrt_config return raw dicts instead),
commit subjects and branch names are passed as arguments instead of through shell
quoting, and mutating operations default to a dry-run preview. See the
MCP Server guide for
Claude Desktop, global install, and HTTP transport with bearer auth.
The MCP server does not cover everything. rrt docs map, rrt docs generate,
rrt docs publish, rrt docs inject, rrt tree --check, rrt toc, rrt changelog lint,
rrt changelog compare, rrt drift generate/check, rrt artifacts --check, every other
--snapshot write, and every rrt-hooks subcommand are CLI-only — a mixed session is
expected, not a fallback.
Prompt phrasings that work
These reliably route to rrt rather than to hand-editing:
"Check with rrt whether this branch name will pass the hooks before you create it."
"Use rrt to preview a minor bump — dry run — and show me every file it would touch."
"Read the Unreleased changelog with rrt before adding an entry, so you don't duplicate one."
"Run rrt doctor and tell me which hook integrations are missing."
"Before you write that commit message, validate the subject with rrt."
"What version is this repo at according to rrt?" — not "what's the version", which invites reading a random file.
"Run rrt release check before you open the PR."
The pattern: name rrt explicitly, and name the moment ("before you create it",
"before you open the PR"). Agents route on triggers, not on capabilities.
Agent instruction snippet
## Use `rrt` for release policy
This repo uses `repo-release-tools` (`rrt`) to enforce branch naming, Conventional
Commits, changelog format, and version consistency. Do not hand-roll any of it.
Before you act, use `rrt`:
| When you are about to… | Use |
|---|---|
| create a branch | `rrt branch new <type> "<desc>"` — or validate the name first with `rrt-hooks check-branch-name --branch <candidate>` |
| write a commit message | `rrt git commit --type <type> "<description>"`, which builds and validates the subject before committing |
| change a version number anywhere | `rrt bump <level> --dry-run` first — never edit version strings by hand; pins and the changelog move with it |
| add a changelog entry | read the existing `[Unreleased]` first; it is hook-managed |
| open a PR | `rrt release check` and `rrt doctor` |
Rules:
- Every mutating `rrt` command takes `--dry-run`. Use it first, show the user the
preview, and only apply after they confirm.
- Never edit a version string by hand in more than one file — that is what `rrt bump` is for.
- Never hand-edit the `[Unreleased]` changelog section while the rrt hooks are active.
- If `rrt` is connected over MCP, prefer the `mcp__rrt__*` tools over shelling out:
typed responses for most tools, no shell quoting of commit subjects, and dry-run is
the default. Shell out for anything with no MCP tool (`rrt docs map`, `rrt docs
generate`, `rrt docs publish`, `rrt docs inject`, `rrt tree --check`, `rrt toc`,
`rrt changelog lint`, `rrt changelog compare`, `rrt drift generate`/`check`,
`rrt artifacts --check`, other `--snapshot` writes, `rrt-hooks *`).
- `rrt --help` lists every command. Check it before concluding rrt cannot do something.Start with the doc that matches your task
Docs index: https://anselmoo.github.io/repo-release-tools/
GitHub Action: https://anselmoo.github.io/repo-release-tools/action/
CLI reference: https://github.com/Anselmoo/repo-release-tools/blob/main/docs/commands/rrt-cli.md
Hook setup: https://github.com/Anselmoo/repo-release-tools/blob/main/docs/commands/hooks.md
Conventional branches: https://github.com/Anselmoo/repo-release-tools/blob/main/docs/commands/branch.md
Git workflow helpers: https://github.com/Anselmoo/repo-release-tools/blob/main/docs/commands/git_cmd.md
Agent skills: https://github.com/Anselmoo/repo-release-tools/blob/main/docs/commands/skill.md
Project tree: https://github.com/Anselmoo/repo-release-tools/blob/main/docs/commands/tree.md
Markdown TOC: https://github.com/Anselmoo/repo-release-tools/blob/main/docs/commands/toc.md
Config health checks: https://github.com/Anselmoo/repo-release-tools/blob/main/docs/commands/doctor.md
Runtime EOL tracking: https://github.com/Anselmoo/repo-release-tools/blob/main/docs/commands/eol_check.md
MCP Server: https://anselmoo.github.io/repo-release-tools/mcp-server/
Agent instructions: https://anselmoo.github.io/repo-release-tools/agent-instructions/
License
repo-release-tools is released under the MIT License.
Some workflow ideas were initially inspired by
joseluisq/gitnow, but the rrt git
surface is intentionally narrower and reshaped around conventional branching,
safe commits, and release automation.
Available Tools
32 toolsgenerate_prefab_uiGenerate Prefab UiA
Execute Prefab Python code in a sandbox and render the result.
The code runs in a Pyodide WASM sandbox with full Python support.
Import everything you use. Use the components tool to look up
available components and their import paths.
Always use PrefabApp as the outermost context manager — this enables streaming so the UI renders progressively as code is written:
from prefab_ui.components import Column, Heading, Text, Row, Badge
from prefab_ui.app import PrefabApp
with PrefabApp() as app:
with Column(gap=4):
Heading("Dashboard")
with Row(gap=2):
Text("Revenue: $1.2M")
Badge("On Track", variant="success")For interactive UIs, pass initial state as a dict and use .rx
on stateful components for reactive bindings:
from prefab_ui.components import Column, Slider, Text
from prefab_ui.app import PrefabApp
with PrefabApp(state={"threshold": 50}) as app:
with Column(gap=4):
slider = Slider(value=50, min=0, max=100, name="threshold")
Text(f"Threshold: {slider.rx}%")slider.rx produces {{ threshold }}, a template expression
that resolves against client-side state. Use Rx("key") directly,
or apply pipe filters: Rx("balance").currency() produces
{{ balance | currency }}.
Available pipes: upper, lower, currency, length, json, round(n), default(val), truncate(n).
Charts live in prefab_ui.components.charts:
from prefab_ui.components.charts import BarChart, ChartSeries
BarChart(
data=[{"month": "Jan", "rev": 100}, {"month": "Feb", "rev": 200}],
series=[ChartSeries(data_key="rev", label="Revenue")],
x_axis="month",
)Values passed via data are available as global variables in the
code. Python features like loops, f-strings, and comprehensions all
work.
Layout patterns:
Card sub-components (CardHeader, CardContent, CardFooter) have built-in padding. Don't add extra padding to them. For a simple card without sub-components, use
Card(css_class="p-6").Use
Grid(columns=N, gap=4)for equal-width cards or panels. Grid handles sizing automatically — no flex classes needed. For unequal widths, pass a list:Grid(columns=[2, 1], gap=4)gives a 2:1 ratio.Row is for inline elements (badges, icons + text, buttons). Prefer Grid when children should have equal or proportional widths. Row does not wrap by default.
Column and Row accept
gap(Tailwind scale: 1-12),align(cross-axis), andjustify(main-axis) as native props — prefer these over raw css_class for spacing.Use
css_class="overflow-hidden"on containers if chart or content edges should clip to the container boundary.
Args: code: Python code that builds a Prefab component tree. data: Values injected as variables in the sandbox namespace. sandbox: A Sandbox instance. If not provided, a new one is created on each call.
Inputs: code (required), data.
rrt_arch_checkDoes the code still match its declared architecture?ARead-onlyIdempotent
Check the tree against [tool.rrt.architecture] and its findings baseline.
Use before adding a module, a unit or a cross-unit import, and before a PR:
the rrt-arch-check hook runs the same check with strict=True. Each finding
is new (not in .rrt/architecture.baseline.toml), baselined, or stale
(recorded but resolved). exit_code matches rrt arch check. Pass
check_templates=True to also self-test every .rrt/templates/ kind. Read-only:
never writes the baseline or renders templates; rrt arch baseline and
rrt arch new stay CLI-only.
Inputs: root, strict, check_templates. Returns: Architecture findings against `.rrt/architecture.baseline.toml`. `exit_code` is…
rrt_artifactsLast recorded artifact-integrity snapshotARead-onlyIdempotent
Return the artifact integrity map from .rrt/artifacts.lock.toml.
Reflects the last rrt artifacts --snapshot, not current state; run the CLI
check to compare against the working tree. If your client can read files
directly, Read .rrt/artifacts.lock.toml is equivalent.
Inputs: no named parameters. Returns: object
rrt_branch_newCreate a policy-compliant branch for this taskADestructive
Derive a compliant branch name from a type + description.
When dry_run=False, checks it out. Use instead of git checkout -b with a
hand-written name — it also returns a matching Conventional Commit title to use
for the first commit, though nothing enforces that the caller actually uses it.
commit_type: a branch type of this repo — the built-ins feat|fix|chore|docs|refactor| test|ci|perf|style|build plus any [[tool.rrt.branch_types]] / extra_branch_types. dry_run=True by default.
Inputs: commit_type (required), description (required), scope, dry_run. Returns: Result of a branch creation (or dry-run preview).
rrt_bumpPreview or apply a release version bumpADestructive
Preview or apply a version bump — use instead of editing version strings by hand.
This is the same pipeline as rrt bump (version targets, pins, changelog
promotion, lockfile and generated-asset refresh, release branch + commit), so a
partial hand-edit will diverge. dry_run=True previews everything and writes
nothing. Each result's changed_paths lists every file the bump touched (or
would touch under dry_run), relative to the repo root.
level: major | minor | patch | alpha | beta | rc. dry_run=True by default.
group: restrict the bump to one [tool.rrt] version group; omit to bump every
configured group (one :class:BumpGroupResult per group either way).
base: patch | minor | major | auto — the core an alpha/beta/rc bump targets
when the current version is FINAL. Same as the CLI's rrt bump rc --base.
Omit it to use the group's prerelease_base config, which defaults to the
next patch (1.0.0 -> 1.0.1-rc.1; minor -> 1.1.0-rc.1; major -> 2.0.0-rc.1).
auto reads Conventional Commits since the last final tag: breaking ->
major, feat -> minor, else patch. Ignored once the version is already a
pre-release, so rc.1 -> rc.2 never moves the core.
scheme: semver | pep440 | calver — the grammar the version is read, bumped
and written in. Same as the CLI's rrt bump --scheme. Omit it to use the
group's version_scheme config; unset there, it is inferred from the
primary target (calendar-shaped -> calver, pep621/python_version -> pep440,
else semver).
Runs the SAME pipeline as the rrt bump CLI command (preflight, version
targets, pin targets, changelog promotion/generation, lockfile and generated-asset
refresh, then release-branch checkout + commit) via
:mod:repo_release_tools.commands.bump's shared stage functions, so an MCP bump
and a CLI bump of the same repo produce identical results (fixes defect D9: the
previous MCP bump only rewrote version-target files and skipped pins, changelog,
lockfiles, generated assets, and git branch/commit entirely).
Inputs: level (required), base, group, scheme, +1 more. Returns: result
rrt_changelogWhat is already in the Unreleased changelog?ARead-onlyIdempotent
Read the changelog.
Use before adding an entry, so you do not duplicate one that is already there,
and before a bump, to see what will be promoted. section='unreleased' (default)
returns parsed pending entries; section='full' returns the raw file. Read-only —
writing an entry needs the CLI or hook workflow (rrt-hooks update-unreleased);
do not hand-edit the [Unreleased] section while that hook is active.
Inputs: section. Returns: result
rrt_configWhat release policy does this repo enforce?ARead-onlyIdempotent
Read the repo's resolved [tool.rrt] policy.
Version targets, pin targets, changelog file, release branch pattern, folder
rules. Call this before proposing any release-related change so you act on
configured policy rather than assumptions. This returns the resolved config as
structured data — a config-load error comes back as a typed ConfigError rather
than a nonzero exit — but it does NOT run the per-target checks rrt config --validate does (target/pin/docs/folder .validate()); use rrt_release_check,
rrt_folder_check, or rrt_docs_check for those.
Inputs: no named parameters. Returns: result
rrt_docs_checkIs the docs lockfile stale?ARead-onlyIdempotent
Check whether .rrt/docs.lock.toml is current against source-owned docs.
Use after editing any source docstring that feeds generated documentation,
before opening a PR — CI fails on this drift. Read-only — never regenerates or
writes files. If stale, run rrt docs generate --format toml to refresh. Covers
.rrt/docs.lock.toml only; rrt docs map --check has no tool here — use the CLI.
Also reports published-docstring skeleton violations in skeleton_issues
when the project configures [tool.rrt.docs.skeleton].
Inputs: no named parameters. Returns: Docs lockfile (.rrt/docs.lock.toml) drift status and skeleton conformance.
rrt_doctorIs this repo's release automation actually wired up?ARead-onlyIdempotent
Check whether hooks, workflows, and the artifact-protection lens are wired correctly.
Covers pre-commit, lefthook, husky, GitHub Actions workflows, and CI artifact-protection. Use before recommending a hook or workflow change, and when a hook did not fire as expected. Returns one CheckResult per component with ok + severity — no output parsing.
Inputs: no named parameters. Returns: result
rrt_doctor_dashboardRRT Doctor DashboardARead-onlyIdempotent
Doctor check results: pass-rate Ring, per-check Metrics, status cards, and detail table.
Renders a UI widget for a human to look at — if you need the values to reason over, call rrt_doctor instead.
Inputs: no named parameters.
rrt_driftLast recorded source-drift snapshotARead-onlyIdempotent
Return source drift state from .rrt/drift.lock.toml (file hashes and symbols).
Reflects the last rrt drift generate, not current state; run rrt drift check
to compare against the working tree. If your client can read files directly,
Read .rrt/drift.lock.toml is equivalent.
Inputs: no named parameters. Returns: object
rrt_eolIs a runtime this repo supports going end-of-life?ARead-onlyIdempotent
Check the host runtime and project minimum versions against EOL policy.
Use before raising or lowering a minimum supported version, and when deciding whether to drop a version from a CI matrix. Read-only — never writes to .rrt/health.lock.toml. Set fetch_live=True to refresh EOL data from endoflife.date instead of the bundled snapshot.
Inputs: language, fetch_live. Returns: Aggregated EOL check results across all requested languages.
rrt_folder_checkDoes this repo's layout violate its own folder policy?ARead-onlyIdempotent
Check the repo layout against [tool.rrt.folders] policy or named built-in templates.
Use before creating a new top-level directory or moving a module, so you place it where policy expects. Read-only — never scaffolds files.
Inputs: template. Returns: Aggregated folder structure policy check results.
rrt_gate_runRun the declared check lines against their expected exitsBDestructive
Run [[tool.rrt.gates]] lines; every line runs and is judged against expect.
Each line returns exit, expected_exit and ok, so a deliberately-red check
that fails as designed reads as ok, and an unexpected failure does not. Use instead
of chaining checks with &&, which stops at the first expected red line.
dry_run=True by default, because the lines are repository-declared commands.
Inputs: names, dry_run. Returns: Result of ``rrt_gate_run``.
rrt_git_commits_sinceHow many commits since a ref, ignoring generated files?ARead-onlyIdempotent
List commits in ..HEAD, leaving out commits that only touch exclude paths.
A file that records HEAD can exclude itself, making "one commit behind" its
resting state. Patterns are root-anchored git globs. exclude_generated adds every
[tool.rrt.git].generated pattern. An unknown ref returns error, never count 0.
Inputs: ref (required), exclude, exclude_generated. Returns: Commits in ``<ref>..HEAD`` with excluded pathspecs left out.
rrt_git_dirty_treeWhich paths are dirty, and which of them are generated?ARead-onlyIdempotent
List every uncommitted path, each marked when it matches [tool.rrt.git].generated.
Use instead of git status before landing or committing. A regenerated manifest
or handoff file shows as generated: true, so you can decide rather than guess.
With ignore_generated=True, clean ignores generated entries. Read-only.
Inputs: ignore_generated. Returns: Result of the dirty-tree check with generated-path awareness.
rrt_git_landLand a finished branch: fast-forward, remove worktree, delete branchADestructive
Fast-forward onto to branch, remove the branch's worktree, and delete it.
One call, one process: three separate git calls cannot do this safely, because a hook can re-dirty the worktree between them. All checks run before any change. Local only — never pushes. Re-running after a failure skips finished steps. restore_generated discards [tool.rrt.git].generated changes in the worktree first. dry_run=True by default.
Inputs: branch (required), onto, dry_run, keep_branch, +1 more. Returns: Result of ``rrt_git_land`` (preview or applied).
rrt_healthLast recorded repo-health snapshotARead-onlyIdempotent
Return the health check results from .rrt/health.lock.toml.
Reflects the last snapshot-writing run, not current state; run the CLI check
to compare against the working tree. If your client can read files directly,
Read .rrt/health.lock.toml is equivalent. This lock can hold results from
three separate commands — rrt doctor --snapshot (pre-commit/lefthook/husky/
workflows), rrt eol --snapshot, and rrt folder --snapshot — so prefer
rrt_doctor, rrt_eol, or rrt_folder_check for a live check of the part you
actually need, not rrt_doctor alone.
Inputs: no named parameters. Returns: object
rrt_health_dashboardRRT Health DashboardARead-onlyIdempotent
Health overview: Metric summary, health Ring, per-lock status chart, check cards, and detail table.
Renders a UI widget for a human to look at — if you need the values to reason over, call rrt_health / rrt_doctor instead.
Inputs: no named parameters.
rrt_initSet up rrt in this repo (interactive form)ARead-onlyIdempotent
Form to initialize rrt configuration — pick target format, preview, then apply.
Renders a UI widget for a human to fill in — if you are an agent working from a
shell, run rrt init directly instead.
Inputs: no named parameters.
rrt_init_runApply rrt init (submitted by the init form)ADestructive
Run rrt init with the given target format. Defaults to dry_run=True for safety.
This is the submit target of the rrt_init form — it shells out to
python -m repo_release_tools.cli init and returns the captured output. If you
are an agent working from a shell, run rrt init directly instead; you gain
nothing by going through this tool.
Inputs: force, target, dry_run. Returns: result
rrt_locks_overviewRRT Locks OverviewARead-onlyIdempotent
All lock files at a glance: status donut chart, Carousel of lock summaries, and full detail table.
Renders a UI widget for a human to look at — if you need the values to reason over, call rrt_health / rrt_tree / rrt_artifacts / rrt_drift instead.
Inputs: no named parameters.
rrt_publish_snapshotForce-push a squashed snapshot to a mirror remote (destructive)ADestructive
Force-push a single-commit snapshot of tracked content to a secondary remote.
dry_run=True by default. The force-push requires BOTH dry_run=False AND
confirm=True -- mirroring the CLI's rrt git publish-snapshot, which
requires an explicit --yes-i-know-this-overwrites-remote-history flag in
addition to not passing --dry-run (two independent signals for one
destructive, history-rewriting operation). If confirm is omitted or False,
the call is always treated as a dry-run preview, regardless of dry_run.
Inputs: remote (required), branch, confirm, dry_run, +2 more. Returns: Result of a publish-snapshot preview or force-push.
rrt_release_checkAre version, pin, and changelog targets consistent for release?ARead-onlyIdempotent
Verify every version target, pin target, and changelog file resolves and agrees, per version group.
Use as the pre-release gate and after any edit that touches a version string in docs or config — pin drift is silent otherwise. Read-only — never modifies files.
Inputs: no named parameters. Returns: Aggregated rrt_release_check results across all configured version groups.
rrt_sync_checkHas upstream released a newer version than we track?ARead-only
List upstream package versions newer than the current project version.
Reads the group's [tool.rrt.upstream] package config and queries the configured registry (PyPI/npm/NuGet/crates.io/Packagist). Read-only — never applies a bump. Requires network access to the registry.
Inputs: group. Returns: Newer upstream package versions available for a version group's tracked package.
rrt_treeLast recorded file-tree snapshot (NOT a live check)ARead-onlyIdempotent
Return the repository tree snapshot from .rrt/tree.lock.toml.
Reflects the last rrt tree --snapshot, not current state — this tool does
NOT run rrt tree --check and cannot tell you whether the working tree still
matches it. There is no MCP tool for that check; use the CLI. If your client
can read files directly, Read .rrt/tree.lock.toml is equivalent.
Inputs: no named parameters. Returns: object
rrt_tree_dashboardRRT Tree DashboardARead-onlyIdempotent
Repository tree: snapshot Metric cards, per-directory bar chart, and clean file table.
Renders a UI widget for a human to look at — if you need the values to reason over, call rrt_tree instead (it is still a snapshot, not a live check).
Inputs: no named parameters.
rrt_validate_branchWill this branch name pass the repo's hooks?ARead-onlyIdempotent
Check a branch name against this repo's configured type allow-list before you create it.
Cheaper than creating it and having the pre-commit hook reject it. Honors [[tool.rrt.branch_types]] and extra_branch_types from config, which a hardcoded regex would miss.
Inputs: branch_name (required). Returns: Result of a branch name validation.
rrt_validate_commitWill this commit subject pass the repo's hooks?ARead-onlyIdempotent
Check a commit subject against Conventional Commits as this repo enforces it, before you commit.
Prefer this over rrt-hooks commit-msg in a shell: the subject is passed as a
string argument, so backticks, quotes, !, and $ in the message cannot be
mangled by shell quoting. Honors [[tool.rrt.commit_types]] and
extra_commit_types from config, as the hook does. Pass branch (the branch
the commit lands on) to also apply that branch type's commit_types
restriction; without it the restriction is not checked.
Inputs: subject (required), branch. Returns: Result of a commit subject validation.
rrt_versionWhat version is this repo at right now?ARead-onlyIdempotent
Read the current version of every configured version group from its primary target.
Use whenever a version number is about to appear in a commit message, changelog heading, docs string, or release note — this is the authoritative value, not whatever a file happened to say. Read-only.
Inputs: no named parameters. Returns: result
rrt_version_overviewRRT Version OverviewARead-onlyIdempotent
Version target map: each configured file, kind, and current version.
Renders a UI widget for a human to look at — if you need the primary-target values to reason over, call rrt_version instead. Unlike this dashboard, which reads every configured target, rrt_version returns only each group's primary target; for secondary/pin target consistency use rrt_release_check.
Inputs: no named parameters.
search_prefab_componentsSearch Prefab ComponentsA
Search the Prefab component library.
Use this tool to look up exact argument names, accepted values, and usage examples before writing component code. The skill covers patterns and layout; this tool has the API details.
The query matches component names and descriptions.
Space-separated terms match independently, so
"Card Badge Metric" returns all three.
When a query matches a small number of components, full details
(docstrings, args, examples) are shown automatically. For broad
searches, a compact listing is returned instead. Use detail to
override this behavior.
Args: query: Filter by component name or description. Space-separated terms are OR-matched. detail: Show full docstrings and args. Defaults to automatic (detailed for ≤5 matches, compact otherwise). limit: Max components to return in detail mode (default 8). No limit in compact mode.
Inputs: limit, query, detail, components. Returns: result
Tool Schema Changelog
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
Loading tool schema changelog…
TDQS
Scored across 32 tools
Descriptions work hard to separate near-siblings (e.g. rrt_health vs rrt_doctor, snapshot readers vs live checks, dashboards vs data tools, and '_dashboard' widgets that explicitly redirect to data tools). A few families remain easy to confuse — rrt_health/rrt_health_dashboard/rrt_locks_overview/rrt_doctor all touch overlapping lock/health data, and rrt_init vs rrt_init_run differ only by 'run'. But overall an agent can tell what to call.
Nearly everything uses a consistent rrt_ snake_case prefix with clear verb/noun semantics (noun reads like rrt_tree/rrt_config, action verbs like rrt_bump/rrt_validate_commit/rrt_git_land). The Prefab tools (generate_prefab_ui, search_prefab_components) break the rrt_ prefix but are a distinct sub-domain. Minor oddity: rrt_init (form) vs rrt_init_run (submit) naming.
32 tools is heavy and includes substantial redundancy: 4 snapshot-lock readers, 4 UI dashboards, plus dashboard-adjacent read tools. For agent use many dashboard widgets duplicate data already available from the reasoning-oriented readers, so several tools do not clearly earn their place.
Covers the full release lifecycle: config, init, version read, bump, changelog, release/doc/folder/eol/arch/sync gates, branch and commit validation, branch creation, git dirty/commits/land/publish, and gate runs. Gaps are explicitly documented as CLI-only (e.g. tree --check, docs map --check) and are minor, so no real dead ends for the stated purpose.
Maintenance
Related MCP Connectors
GitHub Copilot MCP: code, repo, PR, workflow tools for AI assistants.
MCP server that lets your AI assistant build and deploy apps, hosting each one at a live URL.
AI-native git hosting — repos, PRs, issues, CI gates, and AI code review over MCP (60 tools).
MCP server for your apps' tools and custom tools, plus hosted AI agents and approval-gated workflows
Related MCP Servers
- AlicenseAqualityBmaintenanceA MCP server that bridges LLMs with GitHub repository management, enabling automated analysis of pull requests, issue management, tag creation, and release management through natural language.446Apache 2.0
- AlicenseAqualityBmaintenanceA tool that enables AI assistants to conversationally scaffold, build, and publish Python MCP servers to PyPI. It automates the entire development lifecycle, including package naming, tool scaffolding, GitHub repository setup, and package publishing.10MIT
- AlicenseNot gradedqualityDmaintenanceA production-ready MCP server that provides AI assistants with comprehensive GitHub developer tooling including PR analysis, code review, changelog generation, dependency auditing, commit summarization, and refactoring suggestions.7 npmISC
- AlicenseAqualityDmaintenanceMCP server to automate Pull Request creation with AI. Analyzes Git branches, generates descriptions, titles, suggests reviewers, and performs code reviews.84MIT