email-cli-tool 0.1.0__tar.gz → 0.2.0__tar.gz

{email_cli_tool-0.1.0 → email_cli_tool-0.2.0}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: email-cli-tool
-Version: 0.1.0
+Version: 0.2.0
 Summary: A simple CLI tool for sending emails with plain text, HTML, and file attachments
 Project-URL: Homepage, https://github.com/ClaymanTwinkle/email-cli-tool
 Project-URL: Repository, https://github.com/ClaymanTwinkle/email-cli-tool
@@ -123,6 +123,21 @@ Interactively create the configuration file at `~/.emailcli/config.yaml`.
 
 Display current configuration (password is masked).
 
+### `emailcli skill install`
+
+Install the `send-email` agent skill so Claude Code or Codex can send mail for you.
+
+```bash
+# Install for both Claude Code and Codex (default)
+emailcli skill install
+
+# Only one agent
+emailcli skill install --target claude
+emailcli skill install --target codex
+```
+
+Writes `SKILL.md` to `~/.claude/skills/send-email/` and/or `~/.codex/skills/send-email/`.
+
 ## Configuration
 
 Path: `~/.emailcli/config.yaml`

{email_cli_tool-0.1.0 → email_cli_tool-0.2.0}/README.md

@@ -97,6 +97,21 @@ Interactively create the configuration file at `~/.emailcli/config.yaml`.
 
 Display current configuration (password is masked).
 
+### `emailcli skill install`
+
+Install the `send-email` agent skill so Claude Code or Codex can send mail for you.
+
+```bash
+# Install for both Claude Code and Codex (default)
+emailcli skill install
+
+# Only one agent
+emailcli skill install --target claude
+emailcli skill install --target codex
+```
+
+Writes `SKILL.md` to `~/.claude/skills/send-email/` and/or `~/.codex/skills/send-email/`.
+
 ## Configuration
 
 Path: `~/.emailcli/config.yaml`

{email_cli_tool-0.1.0 → email_cli_tool-0.2.0}/README_CN.md

@@ -97,6 +97,21 @@ echo "邮件内容" | emailcli send \
 
 查看当前配置（密码已脱敏）。
 
+### `emailcli skill install`
+
+安装 `send-email` 技能，让 Claude Code 或 Codex 可以直接帮你发邮件。
+
+```bash
+# 默认同时安装到 Claude Code 和 Codex
+emailcli skill install
+
+# 只装其中一个
+emailcli skill install --target claude
+emailcli skill install --target codex
+```
+
+会把 `SKILL.md` 写入 `~/.claude/skills/send-email/` 和/或 `~/.codex/skills/send-email/`。
+
 ## 配置文件
 
 路径：`~/.emailcli/config.yaml`

email_cli_tool-0.2.0/docs/superpowers/plans/2026-06-10-emailcli-skill-install.md

@@ -0,0 +1,497 @@
+# emailcli skill install Implementation Plan
+
+> **For agentic workers:** REQUIRED SUB-SKILL: Use superpowers:subagent-driven-development (recommended) or superpowers:executing-plans to implement this plan task-by-task. Steps use checkbox (`- [ ]`) syntax for tracking.
+
+**Goal:** Add an `emailcli skill install` command that installs the bundled `send-email` agent skill into Claude Code (`~/.claude/skills`) and Codex (`~/.codex/skills`), defaulting to both.
+
+**Architecture:** Ship `SKILL.md` as package data inside `emailcli`. A standalone `skill_install` module reads the bundled file via `importlib.resources` and writes it to per-target user-level directories. `cli.py` adds a thin `skill install` click command that delegates to the module.
+
+**Tech Stack:** Python 3.10+, click, hatchling (build), pytest + click.testing.CliRunner.
+
+---
+
+## File Structure
+
+- **Create** `src/emailcli/skills/send-email/SKILL.md` — canonical skill content (single source of truth, shipped as package data).
+- **Create** `src/emailcli/skill_install.py` — install logic: load bundled content, compute destinations, write per target, return results.
+- **Modify** `pyproject.toml` — force-include the `skills/` data dir into the wheel.
+- **Modify** `src/emailcli/cli.py` — add the `skill` group + `install` subcommand.
+- **Create** `tests/test_skill_install.py` — unit tests for packaging + install logic + CLI command.
+- **Modify** `README.md`, `README_CN.md` — document the command.
+
+Reference spec: `docs/superpowers/specs/2026-06-10-emailcli-skill-install-design.md`
+
+---
+
+### Task 1: Bundle the skill content as package data
+
+**Files:**
+- Create: `src/emailcli/skills/send-email/SKILL.md`
+- Modify: `pyproject.toml:41-42`
+- Test: `tests/test_skill_install.py`
+
+- [ ] **Step 1: Write the failing test**
+
+Create `tests/test_skill_install.py`:
+
+```python
+from importlib.resources import files
+
+
+def test_bundled_skill_md_is_packaged():
+    content = files("emailcli").joinpath("skills/send-email/SKILL.md").read_text(encoding="utf-8")
+    assert content.startswith("---")
+    assert "name: send-email" in content
+    assert "emailcli send" in content
+```
+
+- [ ] **Step 2: Run test to verify it fails**
+
+Run: `uv run pytest tests/test_skill_install.py::test_bundled_skill_md_is_packaged -v`
+Expected: FAIL — `FileNotFoundError` / `IsADirectoryError` (the data file does not exist yet).
+
+- [ ] **Step 3: Create the bundled SKILL.md**
+
+Create `src/emailcli/skills/send-email/SKILL.md` with exactly this content:
+
+````markdown
+---
+name: send-email
+description: "Send emails with attachments via emailcli. Supports plain text, HTML, multiple recipients, and file/image attachments."
+---
+
+# Send Email
+
+Send emails directly from Claude Code using the `emailcli` CLI tool.
+
+## Usage
+
+Invoke with `/send-email` followed by natural language describing what to send:
+
+```
+/send-email 发送测试邮件给 test@example.com
+/send-email 把 report.pdf 发给 alice@example.com，主题是"月度报告"
+/send-email 给 a@example.com 和 b@example.com 发一封 HTML 邮件
+```
+
+## Workflow
+
+### 1. Parse user intent
+
+Extract from the user's message:
+- **Recipients** (`--to`): one or more email addresses
+- **Subject** (`--subject`): email subject line
+- **Body** (`--body` or `--html`): email content
+- **Attachments** (`--attach`): file paths if mentioned
+
+### 2. Validate inputs
+
+- If recipients are missing, ask the user.
+- If subject is missing, ask the user.
+- If body is missing, ask the user.
+- If attachments are mentioned, verify the files exist using Glob or ls before sending.
+
+### 3. Send (no confirmation needed)
+
+Run the `emailcli send` command with the appropriate arguments.
+
+```bash
+emailcli send \
+  --to recipient@example.com \
+  --subject "测试邮件" \
+  --body "Hello World" \
+  --attach report.pdf \
+  --attach photo.png
+```
+
+### 5. Report result
+
+Tell the user whether the email was sent successfully or if there was an error.
+
+## Command Reference
+
+```bash
+# Plain text
+emailcli send --to <email> --subject <subject> --body <text>
+
+# HTML
+emailcli send --to <email> --subject <subject> --html <html>
+
+# HTML from file
+emailcli send --to <email> --subject <subject> --html-file <path>
+
+# With attachments (repeatable)
+emailcli send --to <email> --subject <subject> --body <text> --attach <file>
+
+# Multiple recipients (repeatable)
+emailcli send --to <email1> --to <email2> --subject <subject> --body <text>
+
+# Override sender
+emailcli send --to <email> --subject <subject> --body <text> --from <sender>
+
+# Body from stdin
+echo "content" | emailcli send --to <email> --subject <subject> --body -
+```
+
+## Important
+
+- Do NOT ask for confirmation — send directly once inputs are ready.
+- Verify attachment files exist before running the command.
+- If the command fails, show the error message to the user.
+- The tool must be configured first via `emailcli init`. If sending fails with a config error, tell the user to run `emailcli init`.
+````
+
+- [ ] **Step 4: Force-include the data dir in the wheel**
+
+In `pyproject.toml`, the wheel target currently is:
+
+```toml
+[tool.hatch.build.targets.wheel]
+packages = ["src/emailcli"]
+```
+
+Replace it with:
+
+```toml
+[tool.hatch.build.targets.wheel]
+packages = ["src/emailcli"]
+
+[tool.hatch.build.targets.wheel.force-include]
+"src/emailcli/skills" = "emailcli/skills"
+```
+
+- [ ] **Step 5: Run test to verify it passes**
+
+Run: `uv run pytest tests/test_skill_install.py::test_bundled_skill_md_is_packaged -v`
+Expected: PASS.
+
+- [ ] **Step 6: Verify the wheel actually contains the data file**
+
+Run: `uv build 2>/dev/null && python -c "import zipfile,glob; w=sorted(glob.glob('dist/*.whl'))[-1]; names=zipfile.ZipFile(w).namelist(); print('emailcli/skills/send-email/SKILL.md' in names)"`
+Expected: prints `True`. Then clean up: `rm -rf dist`.
+
+- [ ] **Step 7: Commit**
+
+```bash
+git add src/emailcli/skills/send-email/SKILL.md pyproject.toml tests/test_skill_install.py
+git commit -m "feat: bundle send-email SKILL.md as package data
+
+Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>"
+```
+
+---
+
+### Task 2: Implement the install logic in `skill_install.py`
+
+**Files:**
+- Create: `src/emailcli/skill_install.py`
+- Test: `tests/test_skill_install.py`
+
+- [ ] **Step 1: Write the failing tests**
+
+Append to `tests/test_skill_install.py`:
+
+```python
+from emailcli.skill_install import install_skill, load_skill_content
+
+
+def test_install_all_writes_both_targets(tmp_path):
+    results = install_skill(tmp_path, ["claude", "codex"])
+
+    claude_path = tmp_path / ".claude" / "skills" / "send-email" / "SKILL.md"
+    codex_path = tmp_path / ".codex" / "skills" / "send-email" / "SKILL.md"
+    bundled = load_skill_content()
+
+    assert claude_path.read_text(encoding="utf-8") == bundled
+    assert codex_path.read_text(encoding="utf-8") == bundled
+    assert {r.target for r in results} == {"claude", "codex"}
+    assert all(r.status == "created" for r in results)
+
+
+def test_install_claude_only(tmp_path):
+    install_skill(tmp_path, ["claude"])
+
+    assert (tmp_path / ".claude" / "skills" / "send-email" / "SKILL.md").exists()
+    assert not (tmp_path / ".codex").exists()
+
+
+def test_install_is_idempotent(tmp_path):
+    install_skill(tmp_path, ["claude"])
+    results = install_skill(tmp_path, ["claude"])
+
+    assert results[0].status == "updated"
+    path = tmp_path / ".claude" / "skills" / "send-email" / "SKILL.md"
+    assert path.read_text(encoding="utf-8") == load_skill_content()
+```
+
+- [ ] **Step 2: Run tests to verify they fail**
+
+Run: `uv run pytest tests/test_skill_install.py -k "install" -v`
+Expected: FAIL — `ModuleNotFoundError: No module named 'emailcli.skill_install'`.
+
+- [ ] **Step 3: Write the implementation**
+
+Create `src/emailcli/skill_install.py`:
+
+```python
+from dataclasses import dataclass
+from importlib.resources import files
+from pathlib import Path
+
+SKILL_NAME = "send-email"
+
+# target name -> home subdirectory that holds the agent's skills
+TARGET_HOME_DIRS = {
+    "claude": ".claude",
+    "codex": ".codex",
+}
+
+# human-facing label per target
+TARGET_LABELS = {
+    "claude": "Claude",
+    "codex": "Codex",
+}
+
+
+def load_skill_content() -> str:
+    """Read the bundled SKILL.md shipped as package data."""
+    return (
+        files("emailcli")
+        .joinpath(f"skills/{SKILL_NAME}/SKILL.md")
+        .read_text(encoding="utf-8")
+    )
+
+
+def skill_dest(home: Path, target: str) -> Path:
+    """Destination SKILL.md path for a target under the given home dir."""
+    return home / TARGET_HOME_DIRS[target] / "skills" / SKILL_NAME / "SKILL.md"
+
+
+@dataclass
+class InstallResult:
+    target: str
+    path: Path
+    status: str  # "created" | "updated" | "failed"
+    error: str | None = None
+
+
+def install_skill(home: Path, targets: list[str]) -> list[InstallResult]:
+    """Write the bundled skill into each target's user-level skills dir.
+
+    Each target is handled independently; a write failure on one target is
+    recorded and does not stop the others.
+    """
+    content = load_skill_content()
+    results: list[InstallResult] = []
+    for target in targets:
+        dest = skill_dest(home, target)
+        try:
+            existed = dest.exists()
+            dest.parent.mkdir(parents=True, exist_ok=True)
+            dest.write_text(content, encoding="utf-8")
+            results.append(
+                InstallResult(target, dest, "updated" if existed else "created")
+            )
+        except OSError as exc:
+            results.append(InstallResult(target, dest, "failed", str(exc)))
+    return results
+```
+
+- [ ] **Step 4: Run tests to verify they pass**
+
+Run: `uv run pytest tests/test_skill_install.py -k "install" -v`
+Expected: PASS (3 tests).
+
+- [ ] **Step 5: Commit**
+
+```bash
+git add src/emailcli/skill_install.py tests/test_skill_install.py
+git commit -m "feat: add skill_install module for writing SKILL.md to agent dirs
+
+Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>"
+```
+
+---
+
+### Task 3: Wire up the `skill install` CLI command
+
+**Files:**
+- Modify: `src/emailcli/cli.py` (add imports near `cli.py:1-9`; add new group/command after the `config` group at `cli.py:121-142`)
+- Test: `tests/test_skill_install.py`
+
+- [ ] **Step 1: Write the failing tests**
+
+Append to `tests/test_skill_install.py`:
+
+```python
+from click.testing import CliRunner
+
+from emailcli.cli import cli
+
+
+def test_cli_skill_install_default_writes_both(tmp_path):
+    runner = CliRunner()
+    result = runner.invoke(cli, ["skill", "install", "--home", str(tmp_path)])
+
+    assert result.exit_code == 0
+    assert (tmp_path / ".claude" / "skills" / "send-email" / "SKILL.md").exists()
+    assert (tmp_path / ".codex" / "skills" / "send-email" / "SKILL.md").exists()
+    assert "Claude" in result.output
+    assert "Codex" in result.output
+
+
+def test_cli_skill_install_target_codex(tmp_path):
+    runner = CliRunner()
+    result = runner.invoke(
+        cli, ["skill", "install", "--target", "codex", "--home", str(tmp_path)]
+    )
+
+    assert result.exit_code == 0
+    assert (tmp_path / ".codex" / "skills" / "send-email" / "SKILL.md").exists()
+    assert not (tmp_path / ".claude").exists()
+
+
+def test_cli_skill_install_invalid_target(tmp_path):
+    runner = CliRunner()
+    result = runner.invoke(
+        cli, ["skill", "install", "--target", "vim", "--home", str(tmp_path)]
+    )
+
+    assert result.exit_code != 0
+```
+
+- [ ] **Step 2: Run tests to verify they fail**
+
+Run: `uv run pytest tests/test_skill_install.py -k "cli_skill" -v`
+Expected: FAIL — `No such command 'skill'` (exit code 2 / usage error), so the `exit_code == 0` assertions fail.
+
+- [ ] **Step 3: Add the import**
+
+In `src/emailcli/cli.py`, add to the import block (after the existing `from emailcli.sender import SmtpSender` line):
+
+```python
+from emailcli import skill_install
+```
+
+- [ ] **Step 4: Add the `skill` group and `install` command**
+
+In `src/emailcli/cli.py`, after the `config_group`'s `show` command (end of file), append:
+
+```python
+@cli.group(name="skill")
+def skill_group():
+    """Manage the emailcli agent skill."""
+
+
+@skill_group.command(name="install")
+@click.option(
+    "--target",
+    type=click.Choice(["claude", "codex", "all"]),
+    default="all",
+    show_default=True,
+    help="Which agent to install the skill for.",
+)
+@click.option("--home", default=None, type=click.Path(), hidden=True, help="Home directory (for testing).")
+def install(target, home):
+    """Install the send-email skill into Claude Code and/or Codex."""
+    home_dir = Path(home) if home else Path.home()
+    targets = ["claude", "codex"] if target == "all" else [target]
+
+    results = skill_install.install_skill(home_dir, targets)
+
+    for r in results:
+        label = skill_install.TARGET_LABELS[r.target]
+        if r.status == "failed":
+            click.echo(f"✗ {label}\t{r.path}\t(failed: {r.error})", err=True)
+        else:
+            click.echo(f"✓ {label}\t{r.path}\t({r.status})")
+
+    if any(r.status == "failed" for r in results):
+        raise SystemExit(1)
+```
+
+- [ ] **Step 5: Run tests to verify they pass**
+
+Run: `uv run pytest tests/test_skill_install.py -k "cli_skill" -v`
+Expected: PASS (3 tests).
+
+- [ ] **Step 6: Run the whole test file**
+
+Run: `uv run pytest tests/test_skill_install.py -v`
+Expected: PASS (all tests).
+
+- [ ] **Step 7: Commit**
+
+```bash
+git add src/emailcli/cli.py tests/test_skill_install.py
+git commit -m "feat: add 'emailcli skill install' command
+
+Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>"
+```
+
+---
+
+### Task 4: Document the command
+
+**Files:**
+- Modify: `README.md`
+- Modify: `README_CN.md`
+
+- [ ] **Step 1: Add a section to `README.md`**
+
+In `README.md`, after the `### emailcli config show` block and before `## Configuration`, insert:
+
+````markdown
+### `emailcli skill install`
+
+Install the `send-email` agent skill so Claude Code or Codex can send mail for you.
+
+```bash
+# Install for both Claude Code and Codex (default)
+emailcli skill install
+
+# Only one agent
+emailcli skill install --target claude
+emailcli skill install --target codex
+```
+
+Writes `SKILL.md` to `~/.claude/skills/send-email/` and/or `~/.codex/skills/send-email/`.
+````
+
+- [ ] **Step 2: Add the equivalent section to `README_CN.md`**
+
+In `README_CN.md`, add a matching section in the command reference area (Chinese):
+
+````markdown
+### `emailcli skill install`
+
+安装 `send-email` 技能，让 Claude Code 或 Codex 可以直接帮你发邮件。
+
+```bash
+# 默认同时安装到 Claude Code 和 Codex
+emailcli skill install
+
+# 只装其中一个
+emailcli skill install --target claude
+emailcli skill install --target codex
+```
+
+会把 `SKILL.md` 写入 `~/.claude/skills/send-email/` 和/或 `~/.codex/skills/send-email/`。
+````
+
+- [ ] **Step 3: Commit**
+
+```bash
+git add README.md README_CN.md
+git commit -m "docs: document 'emailcli skill install' command
+
+Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>"
+```
+
+---
+
+## Self-Review Notes
+
+- **Spec coverage:** command interface (Task 3), bundled package data (Task 1), `skill_install` module + per-target independent error handling (Task 2), user-level paths for Claude/Codex (Task 2 `skill_dest`), idempotent overwrite with created/updated reporting (Task 2 + 3), hidden `--home` test hook (Task 3), packaging via force-include + wheel verification (Task 1), docs (Task 4). All spec sections mapped.
+- **Type consistency:** `install_skill(home, targets) -> list[InstallResult]`, `InstallResult(target, path, status, error)`, `load_skill_content()`, `skill_dest()`, `TARGET_LABELS` used identically across module and CLI.
+- **Out of scope (per spec):** no `uninstall`, no project-level install, no Codex legacy prompt format.

email_cli_tool-0.2.0/docs/superpowers/specs/2026-06-10-emailcli-skill-install-design.md

@@ -0,0 +1,122 @@
+# emailcli skill install — Design
+
+Date: 2026-06-10
+
+## Goal
+
+Add a CLI command that installs the `send-email` agent skill into both
+Claude Code and Codex, so users don't have to manually place `SKILL.md`
+into agent directories.
+
+## Background
+
+The `send-email` skill currently lives only as a hand-placed file at
+`~/.claude/skills/send-email/SKILL.md`. It is a thin wrapper that drives the
+`emailcli send` command. There is no skill content shipped inside the package
+and no command to install it.
+
+Claude Code and Codex both load skills from a directory containing a
+`SKILL.md` file with YAML frontmatter (`name`, `description`). The format is
+identical between the two agents, so a single bundled `SKILL.md` can serve
+both — only the destination directory differs.
+
+References:
+- Codex Agent Skills: https://developers.openai.com/codex/skills
+- Claude skills in Codex CLI: https://www.robert-glaser.de/claude-skills-in-codex-cli/
+
+## Scope
+
+- **In scope:** user-level install of the `send-email` skill to Claude and
+  Codex via a new `emailcli skill install` command, default to both targets.
+- **Out of scope:** `uninstall`, project-level install (`./.claude`,
+  `./.codex`), Codex legacy `~/.codex/prompts` format, multi-file skills with
+  `scripts/`/`references/`.
+
+## Decisions
+
+- **Install scope:** user-level only.
+  - Claude: `~/.claude/skills/send-email/SKILL.md`
+  - Codex: `~/.codex/skills/send-email/SKILL.md`
+- **Default target:** both Claude and Codex. `--target` selects one.
+- **Skill source:** bundle `SKILL.md` as package data (single source of
+  truth), read at runtime via `importlib.resources`.
+
+## Command Interface
+
+```bash
+emailcli skill install                 # default: claude + codex
+emailcli skill install --target claude # Claude Code only
+emailcli skill install --target codex  # Codex only
+emailcli skill install --target all    # explicit both
+```
+
+- New click group `skill` with subcommand `install`.
+- `--target` is a `click.Choice(["claude", "codex", "all"])`, default `all`.
+- Idempotent: if the destination `SKILL.md` already exists it is overwritten
+  (this is our managed skill). For each target, print a result line, e.g.
+  `✓ Claude  ~/.claude/skills/send-email/SKILL.md  (updated)` /  `(created)`.
+- Hidden `--home` option (mirrors the existing hidden `--config-dir` pattern)
+  injects a base directory for tests.
+
+## File Layout / Modules
+
+- `src/emailcli/skills/send-email/SKILL.md` — canonical skill content (moved
+  from the existing global copy; the repo becomes the single source of truth).
+- `src/emailcli/skill_install.py` — install logic:
+  - resolve targets from `--target`
+  - compute destination dir per target under a given `home`
+  - read bundled `SKILL.md` via `importlib.resources`
+  - write it, creating parent dirs
+  - return per-target results (target name, path, created-vs-updated)
+  - signature accepts `home: Path` so tests can inject `tmp_path`
+- `src/emailcli/cli.py` — thin click wiring only; delegates to
+  `skill_install`.
+
+## Packaging
+
+Build backend is hatchling, packaging `src/emailcli`. Explicitly ensure the
+skill data is included in the wheel:
+
+```toml
+[tool.hatch.build.targets.wheel.force-include]
+"src/emailcli/skills" = "emailcli/skills"
+```
+
+(or the equivalent `artifacts` glob). Verified during implementation by
+running `uv build` and inspecting the wheel for `emailcli/skills/send-email/SKILL.md`.
+
+Runtime read:
+
+```python
+from importlib.resources import files
+content = files("emailcli").joinpath("skills/send-email/SKILL.md").read_text(encoding="utf-8")
+```
+
+## Error Handling
+
+- Each target is processed independently. If one write fails (e.g.
+  permission error), report that target as failed and continue with the
+  others.
+- If any target failed, exit with a non-zero status; otherwise exit 0.
+- Destination parent directories are created with
+  `mkdir(parents=True, exist_ok=True)`.
+
+## Testing (TDD)
+
+Unit tests for `skill_install` (inject `tmp_path` as `home`):
+- install `all` → both `~/.claude/.../SKILL.md` and `~/.codex/.../SKILL.md`
+  exist and their content equals the bundled `SKILL.md`.
+- `--target claude` → only the Claude file is created; Codex path absent.
+- `--target codex` → only the Codex file is created.
+- running install twice is idempotent (second run reports `updated`, content
+  unchanged).
+
+CLI-level tests via `click.testing.CliRunner` using the hidden `--home`
+option pointed at `tmp_path`:
+- default command writes both targets and prints both result lines.
+- exit code is 0 on success.
+
+## Documentation
+
+Add a short section to `README.md` and `README_CN.md` documenting
+`emailcli skill install` and what it does.

{email_cli_tool-0.1.0 → email_cli_tool-0.2.0}/pyproject.toml

@@ -1,6 +1,6 @@
 [project]
 name = "email-cli-tool"
-version = "0.1.0"
+version = "0.2.0"
 description = "A simple CLI tool for sending emails with plain text, HTML, and file attachments"
 readme = "README.md"
 license = "MIT"

{email_cli_tool-0.1.0 → email_cli_tool-0.2.0}/src/emailcli/cli.py

@@ -7,6 +7,7 @@ from emailcli.config import load_config
 from emailcli.exceptions import EmailCliError
 from emailcli.message import build_message
 from emailcli.sender import SmtpSender
+from emailcli import skill_install
 
 
 @click.group()
@@ -140,3 +141,35 @@ def show(config_dir):
     except EmailCliError as e:
         click.echo(f"Error: {e}", err=True)
         raise SystemExit(1)
+
+
+@cli.group(name="skill")
+def skill_group():
+    """Manage the emailcli agent skill."""
+
+
+@skill_group.command(name="install")
+@click.option(
+    "--target",
+    type=click.Choice(["claude", "codex", "all"]),
+    default="all",
+    show_default=True,
+    help="Which agent to install the skill for.",
+)
+@click.option("--home", default=None, type=click.Path(), hidden=True, help="Home directory (for testing).")
+def install(target, home):
+    """Install the send-email skill into Claude Code and/or Codex."""
+    home_dir = Path(home) if home else Path.home()
+    targets = ["claude", "codex"] if target == "all" else [target]
+
+    results = skill_install.install_skill(home_dir, targets)
+
+    for r in results:
+        label = skill_install.TARGET_LABELS[r.target]
+        if r.status == "failed":
+            click.echo(f"✗ {label}\t{r.path}\t(failed: {r.error})", err=True)
+        else:
+            click.echo(f"✓ {label}\t{r.path}\t({r.status})")
+
+    if any(r.status == "failed" for r in results):
+        raise SystemExit(1)

email_cli_tool-0.2.0/src/emailcli/skill_install.py

@@ -0,0 +1,65 @@
+from dataclasses import dataclass
+from importlib.resources import files
+from pathlib import Path
+from typing import Literal
+
+SKILL_NAME = "send-email"
+
+# target name -> home subdirectory that holds the agent's skills
+TARGET_HOME_DIRS = {
+    "claude": ".claude",
+    "codex": ".codex",
+}
+
+# human-facing label per target
+TARGET_LABELS = {
+    "claude": "Claude",
+    "codex": "Codex",
+}
+
+
+def load_skill_content() -> str:
+    """Read the bundled SKILL.md shipped as package data."""
+    return (
+        files("emailcli")
+        .joinpath(f"skills/{SKILL_NAME}/SKILL.md")
+        .read_text(encoding="utf-8")
+    )
+
+
+def skill_dest(home: Path, target: str) -> Path:
+    """Destination SKILL.md path for a target under the given home dir."""
+    return home / TARGET_HOME_DIRS[target] / "skills" / SKILL_NAME / "SKILL.md"
+
+
+@dataclass
+class InstallResult:
+    target: str
+    path: Path
+    status: Literal["created", "updated", "failed"]
+    error: str | None = None
+
+
+def install_skill(home: Path, targets: list[str]) -> list[InstallResult]:
+    """Write the bundled skill into each target's user-level skills dir.
+
+    Each target is handled independently; a write failure on one target is
+    recorded and does not stop the others.
+    """
+    unknown = [t for t in targets if t not in TARGET_HOME_DIRS]
+    if unknown:
+        raise ValueError(f"Unknown target(s): {', '.join(unknown)}")
+    content = load_skill_content()
+    results: list[InstallResult] = []
+    for target in targets:
+        dest = skill_dest(home, target)
+        try:
+            existed = dest.exists()
+            dest.parent.mkdir(parents=True, exist_ok=True)
+            dest.write_text(content, encoding="utf-8")
+            results.append(
+                InstallResult(target, dest, "updated" if existed else "created")
+            )
+        except OSError as exc:
+            results.append(InstallResult(target, dest, "failed", str(exc)))
+    return results

email_cli_tool-0.2.0/src/emailcli/skills/send-email/SKILL.md

@@ -0,0 +1,84 @@
+---
+name: send-email
+description: "Send emails with attachments via emailcli. Supports plain text, HTML, multiple recipients, and file/image attachments."
+---
+
+# Send Email
+
+Send emails directly from Claude Code using the `emailcli` CLI tool.
+
+## Usage
+
+Invoke with `/send-email` followed by natural language describing what to send:
+
+```
+/send-email 发送测试邮件给 test@example.com
+/send-email 把 report.pdf 发给 alice@example.com，主题是"月度报告"
+/send-email 给 a@example.com 和 b@example.com 发一封 HTML 邮件
+```
+
+## Workflow
+
+### 1. Parse user intent
+
+Extract from the user's message:
+- **Recipients** (`--to`): one or more email addresses
+- **Subject** (`--subject`): email subject line
+- **Body** (`--body` or `--html`): email content
+- **Attachments** (`--attach`): file paths if mentioned
+
+### 2. Validate inputs
+
+- If recipients are missing, ask the user.
+- If subject is missing, ask the user.
+- If body is missing, ask the user.
+- If attachments are mentioned, verify the files exist using Glob or ls before sending.
+
+### 3. Send (no confirmation needed)
+
+Run the `emailcli send` command with the appropriate arguments.
+
+```bash
+emailcli send \
+  --to recipient@example.com \
+  --subject "测试邮件" \
+  --body "Hello World" \
+  --attach report.pdf \
+  --attach photo.png
+```
+
+### 4. Report result
+
+Tell the user whether the email was sent successfully or if there was an error.
+
+## Command Reference
+
+```bash
+# Plain text
+emailcli send --to <email> --subject <subject> --body <text>
+
+# HTML
+emailcli send --to <email> --subject <subject> --html <html>
+
+# HTML from file
+emailcli send --to <email> --subject <subject> --html-file <path>
+
+# With attachments (repeatable)
+emailcli send --to <email> --subject <subject> --body <text> --attach <file>
+
+# Multiple recipients (repeatable)
+emailcli send --to <email1> --to <email2> --subject <subject> --body <text>
+
+# Override sender
+emailcli send --to <email> --subject <subject> --body <text> --from <sender>
+
+# Body from stdin
+echo "content" | emailcli send --to <email> --subject <subject> --body -
+```
+
+## Important
+
+- Do NOT ask for confirmation — send directly once inputs are ready.
+- Verify attachment files exist before running the command.
+- If the command fails, show the error message to the user.
+- The tool must be configured first via `emailcli init`. If sending fails with a config error, tell the user to run `emailcli init`.

email_cli_tool-0.2.0/tests/test_skill_install.py

@@ -0,0 +1,88 @@
+from importlib.resources import files
+from unittest.mock import patch
+
+import pytest
+from click.testing import CliRunner
+
+from emailcli.cli import cli
+from emailcli.skill_install import InstallResult, install_skill, load_skill_content
+
+
+def test_bundled_skill_md_is_packaged():
+    content = files("emailcli").joinpath("skills/send-email/SKILL.md").read_text(encoding="utf-8")
+    assert content.startswith("---")
+    assert "name: send-email" in content
+    assert "emailcli send" in content
+
+
+def test_install_all_writes_both_targets(tmp_path):
+    results = install_skill(tmp_path, ["claude", "codex"])
+
+    claude_path = tmp_path / ".claude" / "skills" / "send-email" / "SKILL.md"
+    codex_path = tmp_path / ".codex" / "skills" / "send-email" / "SKILL.md"
+    bundled = load_skill_content()
+
+    assert claude_path.read_text(encoding="utf-8") == bundled
+    assert codex_path.read_text(encoding="utf-8") == bundled
+    assert {r.target for r in results} == {"claude", "codex"}
+    assert all(r.status == "created" for r in results)
+
+
+def test_install_claude_only(tmp_path):
+    install_skill(tmp_path, ["claude"])
+
+    assert (tmp_path / ".claude" / "skills" / "send-email" / "SKILL.md").exists()
+    assert not (tmp_path / ".codex").exists()
+
+
+def test_install_is_idempotent(tmp_path):
+    install_skill(tmp_path, ["claude"])
+    results = install_skill(tmp_path, ["claude"])
+
+    assert results[0].status == "updated"
+    path = tmp_path / ".claude" / "skills" / "send-email" / "SKILL.md"
+    assert path.read_text(encoding="utf-8") == load_skill_content()
+
+
+def test_install_unknown_target_raises(tmp_path):
+    with pytest.raises(ValueError):
+        install_skill(tmp_path, ["vim"])
+
+
+def test_cli_skill_install_default_writes_both(tmp_path):
+    runner = CliRunner()
+    result = runner.invoke(cli, ["skill", "install", "--home", str(tmp_path)])
+
+    assert result.exit_code == 0
+    assert (tmp_path / ".claude" / "skills" / "send-email" / "SKILL.md").exists()
+    assert (tmp_path / ".codex" / "skills" / "send-email" / "SKILL.md").exists()
+    assert "Claude" in result.output
+    assert "Codex" in result.output
+
+
+def test_cli_skill_install_target_codex(tmp_path):
+    runner = CliRunner()
+    result = runner.invoke(
+        cli, ["skill", "install", "--target", "codex", "--home", str(tmp_path)]
+    )
+
+    assert result.exit_code == 0
+    assert (tmp_path / ".codex" / "skills" / "send-email" / "SKILL.md").exists()
+    assert not (tmp_path / ".claude").exists()
+
+
+def test_cli_skill_install_invalid_target(tmp_path):
+    runner = CliRunner()
+    result = runner.invoke(
+        cli, ["skill", "install", "--target", "vim", "--home", str(tmp_path)]
+    )
+
+    assert result.exit_code != 0
+
+
+def test_cli_skill_install_failure_exit_code(tmp_path):
+    failed = [InstallResult("claude", tmp_path / "SKILL.md", "failed", "boom")]
+    runner = CliRunner()
+    with patch("emailcli.skill_install.install_skill", return_value=failed):
+        result = runner.invoke(cli, ["skill", "install", "--home", str(tmp_path)])
+    assert result.exit_code == 1

{email_cli_tool-0.1.0 → email_cli_tool-0.2.0}/uv.lock

@@ -24,8 +24,8 @@ wheels = [
 ]
 
 [[package]]
-name = "emailcli"
-version = "0.1.0"
+name = "email-cli-tool"
+version = "0.2.0"
 source = { editable = "." }
 dependencies = [
     { name = "click" },