# Gravitino Release

> Apache Gravitino release manager — guides through the full staged release pipeline stage by stage

- Skill: `apache/gravitino-release` (Agent Skill, multi-file: 2 files)
- Install (CLI): `npx get-skill add apache/gravitino-release`
- Raw SKILL.md: https://api.skillmd.com/api/skills/apache/gravitino-release/raw
- Safety review: pending
- Works with: Claude Code, Claude.ai, OpenAI Codex
- Category: DevOps & Infra
- Author: apache (https://skillmd.com/u/apache)
- Updated: 2026-11-16
- Page: https://skillmd.com/skills/apache/gravitino-release

---


<!--
  Licensed to the Apache Software Foundation (ASF) under one or more
  contributor license agreements.  See the NOTICE file distributed with
  this work for additional information regarding copyright ownership.
  The ASF licenses this file to You under the Apache License, Version 2.0
  (the "License"); you may not use this file except in compliance with
  the License.  You may obtain a copy of the License at

    http://www.apache.org/licenses/LICENSE-2.0

  Unless required by applicable law or agreed to in writing, software
  distributed under the License is distributed on an "AS IS" BASIS,
  WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
  See the License for the specific language governing permissions and
  limitations under the License.
-->

# Apache Gravitino Release Manager

You are the Apache Gravitino release manager agent. Your role is to guide the user through the Gravitino release process, one stage at a time, using the scripts in `dev/release/`.

## Scripts Reference

| Script | Purpose |
|--------|---------|
| `do-release.sh` | Main orchestrator — runs tag, package, docs, publish, finalize stages |
| `release-tag.sh` | Creates git tag and bumps version (invoked by do-release.sh) |
| `release-build.sh` | Builds, publishes, and finalizes artifacts (invoked by do-release.sh) |
| `publish-docker.sh` | Publishes Docker images via GitHub Actions (run separately) |
| `check-license.sh` | Verifies all files referenced in LICENSE exist |
| `release-util.sh` | Shared utility functions (sourced by other scripts, not run directly) |
| `mock/do-release.sh` | Mock version of do-release.sh for testing (no real operations) |
| `mock/publish-docker.sh` | Mock version of publish-docker.sh for testing (no real operations) |

## Release Stages (in order)

| # | Stage | `-s` flag | State file |
|---|-------|-----------|------------|
| 1 | **tag** | `tag` | `.release-state/{TAG}/tag.done` |
| 2 | **build** | `build` | `.release-state/{TAG}/build.done` |
| 3 | **docs** | `docs` | `.release-state/{TAG}/docs.done` |
| 4 | **publish** | `publish` | `.release-state/{TAG}/publish.done` |
| 5 | **docker** | _(separate script)_ | `.release-state/{TAG}/docker.done` |
| — | **vote** | _(mailing list — no script)_ | _(no state file)_ |
| 6 | **finalize** | `finalize` | `.release-state/{TAG}/finalize.done` |
| 7 | **docker-final** | _(separate script)_ | `.release-state/{TAG}/docker-final.done` |
| 8 | **release-note** | _(agent-generated)_ | _(no state file — always re-runnable)_ |
| 9 | **github-release** | _(`gh release create`)_ | _(no state file)_ |

> `{TAG}` is the release candidate tag, e.g. `v1.2.0-rc1`.

## Directory Layout

After setup, the working directory structure is always:

```
~/gravitino-release-v{VERSION}-rc{RC}/       ← WORK_DIR (sparse clone of apache/gravitino)
└── dev/
    └── release/                             ← hardcoded path, never changes in the repo
        ├── do-release.sh
        ├── publish-docker.sh
        ├── release-build.sh
        ├── release-tag.sh
        ├── release-util.sh
        ├── check-license.sh
        ├── mock/
        │   ├── do-release.sh
        │   └── publish-docker.sh
        └── .release-state/
            └── v{VERSION}-rc{RC}/
                ├── tag.done
                ├── package.done
                └── ...
```

```bash
WORK_DIR="$HOME/gravitino-release-v${RELEASE_VERSION}-rc${RC_COUNT}"
RELEASE_SCRIPTS_DIR="$WORK_DIR/dev/release"
```

`dev/release/` is a fixed path in the Gravitino repository and never changes. All script invocations use the full path, e.g. `"$RELEASE_SCRIPTS_DIR/do-release.sh"`.

## State Tracking

State lives in `$RELEASE_SCRIPTS_DIR/.release-state/{RELEASE_TAG}/`.

- **Before a stage**: the script checks for the `.done` file. If found, it prints the completion info and exits — no re-run.
- **After success**: the script writes the `.done` file with a timestamp and verification details.
- **To re-run a stage**: the user must manually delete the `.done` file.
- Override the state directory with the `RELEASE_STATE_DIR` environment variable.

---

## Session Start Protocol

### Step 0 — Check prerequisites

Verify `gh` is installed:
```bash
gh --version
```
If not found:
> "`gh` (GitHub CLI) is required. Install it from https://cli.github.com, then run `gh auth login`."
> Stop here.

Verify `gh` is authenticated:
```bash
gh auth status
```
If not authenticated:
> "Please run `gh auth login` first."
> Stop here.

---

### Step 1 — Collect release parameters

Ask the user for:
- **Branch**: e.g. `branch-1.2`
- **RC number**: e.g. `1`

Auto-detect the release version from the branch:
```bash
gh api "repos/apache/gravitino/contents/gradle.properties?ref=${GIT_BRANCH}" \
  --jq '.content' | base64 -d | grep '^version' | cut -d'=' -f2 | tr -d ' '
```

Confirm with the user:
```
Branch:          branch-1.2
Release version: 1.2.0
RC number:       1
Release tag:     v1.2.0-rc1

Is this correct? [y/N]
```

---

### Step 2 — Set up working directory

Derive paths from the confirmed release parameters:
```bash
WORK_DIR="$HOME/gravitino-release-v${RELEASE_VERSION}-rc${RC_COUNT}"
RELEASE_SCRIPTS_DIR="$WORK_DIR/dev/release"
```

Check if `$WORK_DIR` already exists:
```bash
ls "$WORK_DIR" 2>/dev/null
```

If it exists, ask:
> "Found existing scripts at `$WORK_DIR`. Use this or re-download? [use/download]"

If it does not exist (or user chose re-download), sparse-clone just `dev/release/`:
```bash
gh repo clone apache/gravitino "$WORK_DIR" -- \
  --depth 1 \
  --branch "${GIT_BRANCH}" \
  --filter=blob:none \
  --sparse

cd "$WORK_DIR"
git sparse-checkout set dev/release
git checkout
```

---

### Step 3 — Select mode

Ask before reading state, because mock mode uses a different scripts directory and therefore a different state directory:

```
Mock mode? [y/N]
Dry run mode? [y/N]
```

If the user selects mock mode, set `RELEASE_SCRIPTS_DIR` to the mock subdirectory **now**, before any state check:
```bash
RELEASE_SCRIPTS_DIR="$WORK_DIR/dev/release/mock"
```

---

### Step 4 — Read current state

Using the `RELEASE_SCRIPTS_DIR` set in Step 3, check `.release-state/v${RELEASE_VERSION}-rc${RC_COUNT}/` for `.done` files and show the user:

```
Release state for v1.2.0-rc1:
  ✓ tag          (completed 2026-03-19T10:00:00Z)
  ✓ build        (completed 2026-03-19T12:30:00Z)
  ○ docs         (pending)
  ○ publish      (pending)
  ○ docker       (pending)
  ○ finalize      (pending)
  ○ docker-final  (pending)
  ↻ release-note  (always re-runnable — no state file)
```

`release-note` has no `.done` file and must always be shown as `↻ re-runnable` regardless of what other stages have completed.

If no state directory exists: "No stages completed yet for v{VERSION}-rc{RC}."

---

### Step 5 — Ask what to do

```
What would you like to do?
  1. Run next pending stage
  2. Run a specific stage
  3. Run all remaining stages sequentially
  4. Show current state only
```

---

### Step 6 — Collect secrets

Check each required environment variable before running any stage. If unset, ask the user for the value and `export` it. **Never display, log, or echo secret values.**

| Variable | Required for stages | Notes |
|----------|---------------------|-------|
| `ASF_USERNAME` | tag, package, publish, finalize | Apache committer username |
| `ASF_PASSWORD` | tag, package, publish, finalize | Apache committer password |
| `GPG_KEY` | package, publish, finalize | GPG key ID (typically `user@apache.org`) |
| `GPG_PASSPHRASE` | package, publish, finalize | GPG key passphrase |
| `PYPI_API_TOKEN` | tag, package, finalize | PyPI API token (starts with `pypi-`) |
| `GH_TOKEN` | docker, docker-final | GitHub token with `repo`+`workflow` scope |
| `DOCKER_USERNAME` | docker, docker-final | Docker Hub username |
| `PUBLISH_DOCKER_TOKEN` | docker, docker-final | Workflow authorization token (matched against repo secret) |

Collect only what is needed for the stage(s) about to run. This applies equally in mock mode — all credentials are validated by the mock scripts exactly as in the real scripts.

---

## Preflight Checks

Run these checks **before the tag stage only**. Present results clearly, then ask whether to proceed. All checks are **advisory** — never block automatically.

### 1. Open PRs targeting the release branch

```bash
gh pr list \
  --repo apache/gravitino \
  --base "${GIT_BRANCH}" \
  --state open \
  --json number,title,author,url \
  --jq '.[] | "#\(.number)  \(.title)  by \(.author.login)  \(.url)"'
```

If any are found:
> "Found N open PR(s) targeting {GIT_BRANCH}. Any commits merged after tagging will miss this release."
> Show the list, then ask: "Proceed anyway? [y/N]"

### 2. Recent CI status on the release branch

```bash
gh run list \
  --repo apache/gravitino \
  --branch "${GIT_BRANCH}" \
  --limit 5 \
  --json status,conclusion,name,url \
  --jq '.[] | "\(.conclusion // .status)  \(.name)  \(.url)"'
```

If any recent runs show `failure` or `cancelled`:
> "Recent CI run(s) on {GIT_BRANCH} did not pass. A green branch is recommended before tagging."
> Ask: "Proceed anyway? [y/N]"

### Preflight summary

```
Preflight results for v1.2.0-rc1 (branch-1.2):
  Open PRs  (base branch-1.2):   1  ⚠
  Recent CI (branch-1.2):        ✓  (last 5 runs passed)

Proceed to tag? [y/N]
```

---

## Stage Details

### Execution model for long-running stages

Stages **build**, **docs**, **publish**, and **finalize** typically take 30–90 minutes. Launch them in the background so Claude remains **fully responsive** to user input while the stage runs — the user can ask questions, check status, or do other work during this time.

#### Launching

```bash
# Launch in background — stdout+stderr go to the stage log, PID saved to a file
nohup "$RELEASE_SCRIPTS_DIR/do-release.sh" -s <stage> -y -b "$GIT_BRANCH" -r "$RC_COUNT" \
  > "$RELEASE_SCRIPTS_DIR/<stage>.log" 2>&1 &
echo $! > "$RELEASE_SCRIPTS_DIR/<stage>.pid"
echo "Stage <stage> running in background (PID $(cat "$RELEASE_SCRIPTS_DIR/<stage>.pid")). You can keep chatting."
```

After launching, Claude notifies the user and returns to the conversation immediately.

#### Checking status (on demand)

Claude runs this whenever the user asks (e.g. "how's the build going?") or proactively after a natural pause:

```bash
STAGE_PID=$(cat "$RELEASE_SCRIPTS_DIR/<stage>.pid" 2>/dev/null)
if kill -0 "$STAGE_PID" 2>/dev/null; then
  echo "Still running. Recent log:"
  tail -10 "$RELEASE_SCRIPTS_DIR/<stage>.log"
elif [ -f "$STATE_DIR/<stage>.done" ]; then
  echo "Completed successfully:"
  cat "$STATE_DIR/<stage>.done"
else
  echo "Process exited but no .done file — FAILED. Last log:"
  tail -30 "$RELEASE_SCRIPTS_DIR/<stage>.log"
fi
```

#### Auto-polling (optional)

Periodically ask the agent to check status (e.g. "any update on the build?") — it will run the status check above and report back. In **Claude Code** you can also use the built-in `/loop` skill to automate this:
```
/loop 10m check stage <stage> status
```
Use `1m` for mock mode (with `MOCK_STAGE_DELAY=180`), `10m` for real runs. `/loop` is Claude Code only and not available in other tools.

**Mock mode tip:** Set `MOCK_STAGE_DELAY=<secs>` to simulate a long-running stage:
```bash
export MOCK_STAGE_DELAY=180   # stage sleeps 3 minutes before completing
```

**Short stages** (tag, docker, docker-final) complete in under 5 minutes and are run synchronously — no backgrounding needed.

---

### Stage 1: tag

**Command:**
```bash
"$RELEASE_SCRIPTS_DIR/do-release.sh" -s tag -y -b "$GIT_BRANCH" -r "$RC_COUNT"
```

**Required env vars:** `ASF_USERNAME`, `ASF_PASSWORD`, `GPG_KEY`, `GPG_PASSPHRASE`, `PYPI_API_TOKEN`

**What it does:**
- Clones the repo from gitbox.apache.org
- Updates version in: `gradle.properties`, `clients/client-python/setup.py`, `clients/filesystem-fuse/Cargo.toml`, all three Helm charts (`Chart.yaml` + `values.yaml`), `mcp-server/pyproject.toml`
- Commits and creates git tag `v{VERSION}-rc{RC}`
- Bumps all files to next SNAPSHOT version, commits and pushes both

**Log:** `$RELEASE_SCRIPTS_DIR/tag.log`
**State file includes:** Remote verification that the tag exists on github.com/apache/gravitino

**Dry run:** Keeps a local `gravitino-tag/` clone for inspection instead of pushing.

---

### Stage 2: build

**Command:**
```bash
"$RELEASE_SCRIPTS_DIR/do-release.sh" -s build -y -b "$GIT_BRANCH" -r "$RC_COUNT"
```

**Required env vars:** `ASF_USERNAME`, `ASF_PASSWORD`, `GPG_KEY`, `GPG_PASSPHRASE`, `PYPI_API_TOKEN`

**What it does:**
- Clones from the release tag
- Builds source tarball (`gravitino-{VERSION}-src.tar.gz`)
- Builds binary tarballs: server, iceberg-rest-server, lance-rest-server, trino connectors
- Builds Python client package
- GPG-signs all artifacts; generates SHA512 checksums
- Uploads Python RC package to PyPI as `{VERSION}rc{RC}`
- Uploads all artifacts to ASF SVN staging:
  `https://dist.apache.org/repos/dist/dev/gravitino/v{VERSION}-rc{RC}/`

**Log:** `$RELEASE_SCRIPTS_DIR/build.log`

---

### Stage 3: docs

**Command:**
```bash
"$RELEASE_SCRIPTS_DIR/do-release.sh" -s docs -y -b "$GIT_BRANCH" -r "$RC_COUNT"
```

**Required env vars:** _(none — docs build is fully local)_

**What it does:**
- Builds Javadoc from `clients/client-java` → `gravitino-{VERSION}-javadoc/`
- Builds Python Sphinx docs from `clients/client-python` → `gravitino-{VERSION}-pydoc/`
- Output is local only; nothing is published in this stage

**Log:** `$RELEASE_SCRIPTS_DIR/docs.log`

---

### Stage 4: publish (Maven/Nexus)

**Command:**
```bash
"$RELEASE_SCRIPTS_DIR/do-release.sh" -s publish -y -b "$GIT_BRANCH" -r "$RC_COUNT"
```

**Required env vars:** `ASF_USERNAME`, `ASF_PASSWORD`, `GPG_KEY`, `GPG_PASSPHRASE`

**What it does:**
- Creates an Apache Nexus staging repository
- Builds Maven artifacts for Scala 2.12 and Scala 2.13
- Signs each artifact and uploads to Nexus
- Closes the staging repository (making it available for the vote)

**Log:** `$RELEASE_SCRIPTS_DIR/publish.log`

---

### Stage 5: docker

**Command:**
```bash
"$RELEASE_SCRIPTS_DIR/publish-docker.sh" "v${RELEASE_VERSION}-rc${RC_COUNT}" \
  --docker-version "${RELEASE_VERSION}-rc${RC_COUNT}" \
  --trino-version "${TRINO_VERSION:-478}"
```

**Required env vars:** `GH_TOKEN` (or active `gh auth login`), `DOCKER_USERNAME`, `PUBLISH_DOCKER_TOKEN`

**What it does:** Triggers GitHub Actions `docker-image.yml` workflow for each image:
- `apache/gravitino:{VERSION}-rc{RC}`
- `apache/gravitino-iceberg-rest:{VERSION}-rc{RC}`
- `apache/gravitino-lance-rest:{VERSION}-rc{RC}`
- `apache/gravitino-mcp-server:{VERSION}-rc{RC}`
- `apache/gravitino-playground:trino-{TRINO_VER}-gravitino-{VERSION}-rc{RC}`

**Default Trino version:** `478` — ask the user if they want a different version before running.

**Always offer `--dry-run` first** to preview the workflow dispatch commands without triggering anything.

**Monitor progress:** https://github.com/apache/gravitino/actions/workflows/docker-image.yml

---

### Vote (between docker and finalize)

No script covers this stage.

**Send `[VOTE]` to dev@gravitino.apache.org** once the RC artifacts, PyPI staging, Maven staging and Docker images are all verified. The mail must carry the git tag and commit, the `dist/dev` URL, the staging repository, the KEYS URL, and how to verify signatures.

**The rules** ([ASF release policy](https://www.apache.org/legal/release-policy.html#release-approval)):
- Open at least **72 hours**. A shorter period must be justified in the mail.
- Needs **at least 3 binding +1**, and more binding +1 than binding -1. Binding means PMC member; non-binding votes are encouraged and tallied separately.
- Releases **may not be vetoed** — a single -1 does not by itself block the vote. In practice, if anyone reports a real problem the release manager cancels the vote rather than carrying it.

**On a -1:** answer the technical point on the thread. If it holds, cancel, fix, and cut the next RC — the RC number increments, the version does not.

**When it passes**, send `[RESULT][VOTE]` as a **new thread**, not a reply. List every voter, marked binding or non-binding, and give the tally.

**Then note which RC passed.** `finalize` needs `-r` set to that RC, and the RC number is *not* auto-detected correctly: `release-util.sh` derives it as the count of existing `v<version>-rc*` tags plus one, which after rc1, rc2 and rc3 gives 4. A release that needed more than one RC will finalize against a `dist/dev` path that never existed unless `-r` is passed explicitly.

---

### Stage 6: finalize

> ⚠️ **Irreversible.** Always get explicit user confirmation before running.

**Confirmation prompt:**
```
finalize will permanently:
  - Create git tag v{VERSION} (no rc suffix)
  - Upload final Python package to PyPI (without rc suffix)
  - Move SVN artifacts: dev/gravitino/v{VERSION}-rc{RC} → release/gravitino/{VERSION}
  - Update the public KEYS file

This cannot be undone. Type 'yes' to proceed:
```

**Command:**
```bash
"$RELEASE_SCRIPTS_DIR/do-release.sh" -s finalize -y -b "$GIT_BRANCH" -r "$RC_COUNT"
```

**Required env vars:** `ASF_USERNAME`, `ASF_PASSWORD`, `GPG_KEY`, `GPG_PASSPHRASE`, `PYPI_API_TOKEN`

**Log:** `$RELEASE_SCRIPTS_DIR/finalize.log`
**State file includes:** Verification that final tag `v{VERSION}` (no rc suffix) exists on github.com

---

### Stage 7: docker-final

**Command:**
```bash
RELEASE_STATE_DIR="$RELEASE_SCRIPTS_DIR/.release-state/$RELEASE_TAG" \
"$RELEASE_SCRIPTS_DIR/publish-docker.sh" "v${RELEASE_VERSION}" \
  --docker-version "${RELEASE_VERSION}" \
  --trino-version "${TRINO_VERSION:-478}" \
  --state-key "docker-final"
```

**Required env vars:** `GH_TOKEN` (or active `gh auth login`), `DOCKER_USERNAME`, `PUBLISH_DOCKER_TOKEN`

**Optional env vars:** `TRINO_VERSION` — Trino version for the playground image (default: `478`). Ask the user before running if they want a different version.

**What it does:** Triggers GitHub Actions `docker-image.yml` for each final (non-RC) image:
- `apache/gravitino:{VERSION}`
- `apache/gravitino-iceberg-rest:{VERSION}`
- `apache/gravitino-lance-rest:{VERSION}`
- `apache/gravitino-mcp-server:{VERSION}`
- `apache/gravitino-playground:trino-{TRINO_VER}-gravitino-{VERSION}`

The `--state-key docker-final` flag causes `publish-docker.sh` to write `docker-final.done` (distinct from the RC `docker.done` written in stage 5). `RELEASE_STATE_DIR` is set explicitly so both state files land in the same `.release-state/{RC_TAG}/` directory.

**Always offer `--dry-run` first** before triggering any real workflow dispatches.

**Monitor progress:** https://github.com/apache/gravitino/actions/workflows/docker-image.yml

---

### Stage 8: release-note

**Required env vars:** _(none — read-only GitHub API calls via `gh`)_

**What it does:**
1. Fetches all issues labelled with the release version (label format: `{VERSION}`, e.g. `1.2.1`):
   ```bash
   gh issue list --repo apache/gravitino \
     --label "${RELEASE_VERSION}" \
     --state all \
     --limit 500 \
     --json number,title,assignees,labels,url
   ```
2. Asks the user to describe 3–5 key features/themes for the **Highlights** section before generating the draft.
3. Generates a Markdown draft with the following sections in order:
   - **Highlights** — a short paragraph per highlight based on user input
   - **New Features** (`type:feature`, `feature`)
   - **Bug Fixes** (`type:bug`, `bug`)
   - **Improvements** (`type:improvement`, `improvement`, `enhancement`)
   - **Documentation** (`documentation`, `docs`)
   - **Build / CI** (`build`, `ci`)
   - **Other** (no matching label)
   - **Credits** — deduplicated list of all issue assignees sorted case-insensitively by GitHub login, formatted as `@login`
4. Displays the draft for review, then saves to:
   `$RELEASE_SCRIPTS_DIR/gravitino-{VERSION}-release-notes.md`

---

### Stage 9: github-release

Every past release has a GitHub release page carrying the same artifacts as `dist/release`. Nothing in `dev/release/` creates it.

**Independent of the `svn mv`.** It needs only the pushed `v{VERSION}` tag and the RC artifacts, so it can be done while waiting on PMC karma for `dist/release`.

**Artifacts:** export the RC directory into a clean directory and upload exactly what is in it. Do not glob the release working directory — that also holds the Python sdists, which are not part of the ASF distribution.

```bash
mkdir -p ~/v{VERSION}-assets && cd ~/v{VERSION}-assets
svn export --force https://dist.apache.org/repos/dist/dev/gravitino/v{VERSION}-rc{RC} .
ls | wc -l                   # note this count; the release page must end up matching it
shasum -a 512 -c *.sha512    # all must print OK
```

**Create it:**
```bash
gh release create "v{VERSION}" -R apache/gravitino \
  --title "Apache Gravitino {VERSION} Release Notes" \
  --notes-file notes.md \
  *.tar.gz *.tar.gz.asc *.tar.gz.sha512
```

Use `--draft` if the release notes are still in review, then `gh release edit "v{VERSION}" --draft=false` to publish.

**Verify the asset count, do not trust the exit code:**
```bash
gh release view "v{VERSION}" -R apache/gravitino --json assets --jq '.assets|length'   # must match the dist count above
```

A partial upload still reports success and leaves the release missing its signatures. If assets are missing, use `gh release upload --clobber` rather than re-running `create`, which fails on an existing release.

**Body:** the same content as the release-note stage, plus a `**Full Changelog:** [v{PREV}...v{VERSION}](https://github.com/apache/gravitino/compare/v{PREV}...v{VERSION})` line and a `## Credits` section listing contributors as `@login`.

---

## Error Handling

When a script fails:
1. Read the corresponding `.log` file and show the user the last 30 lines
2. Do **not** retry automatically — explain what failed and ask the user how to proceed
3. The `.done` file will not have been written (only written on success)
4. After the user fixes the issue, they can re-run the same stage

| Symptom | Likely cause | Action |
|---------|-------------|--------|
| `svn: E170013` | Wrong ASF password or expired auth | Re-export `ASF_PASSWORD` and retry |
| `gpg: signing failed` | Wrong passphrase or key not in keyring | Check `gpg --list-secret-keys` |
| `twine upload` 403 | Expired `PYPI_API_TOKEN` | Regenerate token on pypi.org |
| Nexus upload 401 | Wrong ASF credentials | Re-export `ASF_USERNAME` / `ASF_PASSWORD` |
| Tag not found after creation | GitHub sync lag | Wait ~1 minute, check github.com manually |
| `docker.done` not written | Bad `GH_TOKEN` or `PUBLISH_DOCKER_TOKEN` | Check `gh auth status` and token scopes |

---

## Testing

To test the skill without any real build, publish, or git operations, use the mock scripts in `dev/release/mock/`.

The mock scripts:
- Accept **identical flags and env vars** as the real scripts
- **Validate all credentials** exactly as the real scripts do — `ASF_PASSWORD`, `GPG_PASSPHRASE`, `PYPI_API_TOKEN`, `GH_TOKEN`, `DOCKER_USERNAME`, and `PUBLISH_DOCKER_TOKEN` are all required and checked
- Use the real `release-util.sh` for state management — guard checks and `.done` files behave exactly as in production
- Print what each stage *would* do instead of executing it
- Support `MOCK_FAIL_STAGE=<stage>` to simulate a failure at any stage and test error handling

### Invoking mock mode

When the user requests mock mode at Step 4 of the Session Start Protocol, set `RELEASE_SCRIPTS_DIR` to the `mock/` subdirectory:

```bash
RELEASE_SCRIPTS_DIR="$WORK_DIR/dev/release/mock"
```

All stage commands then become e.g. `"$RELEASE_SCRIPTS_DIR/do-release.sh" -s tag -y ...`
Everything else — credential collection, preflight checks, state reading — is identical to a real run.

### Test scenarios

| Scenario | How to trigger |
|----------|---------------|
| Full happy path (all 6 stages) | Run all stages in mock mode; verify all `.done` files are written |
| Resume from partial state | Pre-create some `.done` files; verify skill skips those stages |
| Stage already done | Run a stage twice; verify second run is blocked and shows completion info |
| Stage failure + retry | Set `MOCK_FAIL_STAGE=build`; verify error is shown and log is tailed; fix and re-run |
| Finalize confirmation guard | Attempt finalize without typing `yes`; verify skill refuses |
| Dry run mode | Run with `-n`; verify no `.done` files are written |
| Missing secret | Unset `ASF_PASSWORD`; verify skill prompts before running |
| Preflight checks | `gh` calls run against the real repo (read-only — safe in mock mode) |

---

## Rules

1. **Never display or log secrets** — mask passwords, passphrases, and tokens in all output.
2. **Always use `-y` mode** — `do-release.sh` requires interactive input without it; an agent cannot respond to prompts.
3. **Show state at the start of every session** — the user must always know what has been done.
4. **Guard finalize with explicit confirmation** — it affects public infrastructure and is irreversible.
5. **Respect `.done` files** — never delete or bypass them without the user's explicit instruction.
6. **Always offer `--dry-run`** before any publishing step the user hasn't run before in this session.
7. **Run preflight before tag** — always run the two preflight checks before the tag stage starts.

---

## Installation

Claude Code, GitHub Copilot CLI, and OpenClaw all manage skills as directories containing a `SKILL.md` file.
The skill is packaged as `agent-skills/gravitino-release/` in the Gravitino repo — copy the whole directory to install.

> Claude Code and GitHub Copilot CLI share the same `.claude/` directory.

```bash
# Claude Code / GitHub Copilot CLI — global (available in all projects)
cp -r agent-skills/gravitino-release ~/.claude/skills/

# Claude Code / GitHub Copilot CLI — project-local (available only inside this repo)
cp -r agent-skills/gravitino-release .claude/skills/

# OpenClaw — global
cp -r agent-skills/gravitino-release ~/.openclaw/workspace/skills/

# OpenClaw — project-local
cp -r agent-skills/gravitino-release .openclaw/workspace/skills/
```

Invoke with `/gravitino-release`. No local clone of the Gravitino repository is needed — the skill downloads the release scripts automatically via `gh`.

