Slash Command

/podcast

From podcast-creator

Turn a source (topic, URL, PDF, markdown, repo) into a finished branded podcast episode, locally.

Invocation

How this command is triggered — by the user, by Claude, or both

Slash command

/podcast-creator:podcast <source> [--show <profile>] [--out <dir>]

Model invocable

No pre-commands

Context Preview

The summary Claude sees in its command listing — used to decide when to auto-load this command

# /podcast

Enter the **podcast-studio** orchestrator and produce a local episode from a
source.

Invoke the `podcast-creator:podcast-studio` skill and run its workflow against
the arguments below. Do not produce audio yourself — drive the skill.

## Arguments

`$ARGUMENTS`

Parse them as:

- `<source>` (first positional) — what to make the episode from: a topic
  string, a URL, a path to a mounted/dropped PDF or markdown file, or a repo.
  If no source is given, ask the user for one (one line) before proceeding.
- `--show <profile>` — the show profile to apply (house tone, language,
  off-...

Command Content

54 lines · ~584 tokens

Stats

LanguagePython

Parent stars0

MaintenanceExcellent

Last CommitJun 26, 2026

Actions

View Source View Plugin View on GitHub View README

Stats

Actions

/podcast

Enter the podcast-studio orchestrator and produce a local episode from a source.

Invoke the podcast-creator:podcast-studio skill and run its workflow against the arguments below. Do not produce audio yourself — drive the skill.

Arguments

$ARGUMENTS

Parse them as:

<source> (first positional) — what to make the episode from: a topic string, a URL, a path to a mounted/dropped PDF or markdown file, or a repo. If no source is given, ask the user for one (one line) before proceeding.
--show <profile> — the show profile to apply (house tone, language, off-limits, brand-name rule, default format). Defaults to the PODCAST_CREATOR_SHOW env var, then default.
--out <dir> — output directory for deliverables. Defaults to ./podcast-output/<slug>/.

What to do

Follow the podcast-studio SKILL.md: resolve the show profile, resolve the Gemini credential via the local loader (the sole key source — never print or ask for it in chat), and resolve the output directory.
Run the 7-step pipeline (research → script → TTS → mix → metadata → cover → finish), injecting the resolved key into pipeline subprocesses as GEMINI_API_KEY (in-memory env only). Per the SKILL.md Step 0, resolve the uv-managed venv interpreter once ($PODCAST_PY) and run every dep-needing pipeline script (and check_run.py) through it — never bare python3 (R70).
If a required pipeline skill or profile directory is genuinely missing from the plugin, or the credential store is missing, say so plainly and stop — do not fabricate.

First-run / credential setup runs outside this conversation:

python3 "${CLAUDE_PLUGIN_ROOT}/skills/podcast-studio/scripts/setup_credentials.py" [profile]

Dependency / environment readiness is not a script — on first run the orchestrator reads skills/podcast-studio/knowledge/onboarding-rubric.md, probes the environment with Bash (python3, the pip libs, ffmpeg, the Gemini key presence, network, source, profile, music kit), and installs what it can with your confirmation. The only step you run yourself is the credential setup above.