hardhackerlabs/podwise-cli / skills
podwise
FlaggedPodcast knowledge workflows powered by Podwise CLI: search podcasts and episodes by keyword, monitor followed shows for new releases, find popular episodes, ask questions and extract insights from transcript content, process Podwise episode URLs, YouTube videos, Xiaoyuzhou links, and local audio or video files to retrieve transcripts, summaries, chapters, Q&A, mind maps, highlights, and keywords — plus catch up on your backlog, refine your listening taste, generate weekly recaps, export episode notes to PKM tools, research topics across podcasts, debate episode ideas, and generate language learning cards. Use when the user wants to find, summarize, transcribe, or extract insights from any podcast or audio content, or manage their listening library.
| File | Kind | Size |
|---|---|---|
| SKILL.md | skill md | 5.4 KB |
| references/cli.md | markdown | 10.3 KB |
| references/installation.md | markdown | 2.1 KB |
| workflows/catch-up.md | markdown | 6.0 KB |
| workflows/discover.md | markdown | 8.3 KB |
| workflows/episode-debate.md | markdown | 8.4 KB |
| workflows/episode-notes.md | markdown | 7.7 KB |
| workflows/language-learning.md | markdown | 8.8 KB |
| workflows/refine-taste.md | markdown | 5.4 KB |
| workflows/topic-research.md | markdown | 7.4 KB |
| workflows/weekly-recap.md | markdown | 8.7 KB |
workflows/refine-taste.md
View on GitHub# Podwise Refine Taste
Use this skill to construct a `taste.md` file that captures the user's podcast taste, topic interests, listening habits, and preferred output style. This file is read by other Podwise skills to personalize their outputs.
## Goals
1. Verify that `podwise` is installed and configured.
2. Collect the user's subscriptions and recent listening activity using the CLI.
3. Ask a small set of clarifying questions to fill in things the CLI cannot infer.
4. Write a `taste.md` file to the working directory.
5. Confirm where the file was saved and tell the user which skills will use it.
## Step 1: Check the Environment
```bash
podwise --help
podwise config show
```
If `podwise` is not installed or the API key is missing, stop and follow [references/installation.md](references/installation.md) before continuing.
## Step 2: Gather Data from the CLI
Run these commands to collect raw material for the taste profile:
```bash
# All podcasts the user follows (podcasts with recent updates as proxy for subscription list)
podwise list podcasts --json --latest 30
# Real listening behaviour (primary signal for engagement)
podwise history listened --json --limit 100
# Read activity (exploration behaviour beyond regular listens)
podwise history read --json --limit 100
```
Parse the output to extract:
- Podcast names, categories, and languages (from `list podcasts`)
- Which shows appear most frequently in `history listened` (signals strong engagement)
- Which shows are subscribed but never appear in history (signals low engagement — candidates for Deprioritize)
- Episode titles and topics from `history read` to understand what the user explores beyond their regular listens
## Step 3: Ask Clarifying Questions
The CLI data shows *what* the user subscribes to but not *why* or *how* they listen. Ask the following questions — all at once in a single message, not one by one:
1. **Topics you care about most**: Of the podcasts you follow, which topics or themes matter most to you right now? (e.g. AI, investing, language learning, startup building)
2. **Shows you actually listen to**: Are there shows you follow but rarely listen to, and shows you never miss? If so, which ones?
3. **Episode style preference**: Do you prefer long-form deep dives, short daily news, or interview-style conversations?
4. **Language setup**: Do you listen in multiple languages? If so, which is your primary learning language (if any)?
5. **What you do with insights**: Do you take notes, export to a PKM tool, or just listen? Which tool if any? (Notion, Obsidian, Logseq, Readwise)
6. **Output tone preference**: When Podwise AI summarizes or recommends things for you, do you prefer concise bullet points, flowing prose, or a mix?
## Step 4: Write taste.md
Synthesize the CLI data and the user's answers into a structured file. Save it as `taste.md` in the current working directory unless the user specifies another path.
Use this template:
```markdown
# Listener Taste
_Last updated: {date}_
## Subscribed Podcasts
{List each podcast with name, category/topic, language, and a one-line note on how active the user is with it — inferred from recency of episodes in their feed.}
## Core Interest Areas
{2–5 topic areas the user cares about most, ranked by apparent priority. Derived from podcast categories + user's own answers.}
## Listening Style
- **Episode format preference**: {long-form / short daily / interview / mix}
- **Languages**: {primary language} {+ learning language if applicable}
- **Typical listening context**: {inferred or stated — e.g. commute, focused study, background}
## What the User Does with Insights
- **Note-taking**: {yes / no}
- **PKM tool**: {Notion / Obsidian / Logseq / Readwise / none}
- **Export habits**: {e.g. saves highlights, exports weekly recap, rarely exports}
## Output Preferences
- **Preferred format**: {bullet points / prose / mix}
- **Preferred summary length**: {short / medium / detailed}
- **Preferred language for AI output**: {language}
## Shows to Prioritize
{3–5 shows the user identified as ones they never miss. These should be weighted higher in catch-up, recap, and discovery outputs.}
## Shows to Deprioritize
{Shows the user follows but rarely engages with. Catch-up and recap should surface these last.}
```
Fill every section. If a field cannot be inferred and the user did not answer it, write `unknown` — do not omit the field.
## Step 5: Confirm and Summarize
After writing the file:
1. Print the full path where `taste.md` was saved.
2. Show the user a brief summary of what was captured: number of subscribed shows, core interest areas, and PKM tool detected.
3. Tell the user which skills will read this file automatically: `podwise-catch-up`, `podwise-weekly-recap`, `podwise-discover`, `podwise-topic-research`, and `podwise-episode-debate`.
4. Suggest re-running this skill if their listening habits change significantly.
## Common Failure Cases
- If `podwise list podcasts` returns an empty list, the user has no subscriptions yet. Ask them to follow at least a few podcasts in Podwise before running this skill.
- If the user skips the clarifying questions, write the taste profile from CLI data alone and mark skipped fields as `unknown`.
- If `taste.md` already exists, ask whether to overwrite or append before writing.
## Output Contract
Produce exactly one file: `taste.md`.
Do not print the full file contents in the conversation unless the user asks to review it — instead show the summary and the file path.
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| add-product-e2e-evalpaperclipai/paperclipAdd or extend a Paperclip full-stack runner E2E workflow, fixture, matcher, or report evidence path for local or Daytona execution. | Stars: 115k | Updated | Claude CodeCodex | Verified | Install |
| add-runner-evalpaperclipai/paperclipAdd or extend a Paperclip Runner protocol evaluation definition, roster, assertion, or report fixture with provenance and narrow validation. | Stars: 115k | Updated | Claude CodeCodex | Verified | Install |
| diagnose-why-work-stoppedpaperclipai/paperclipCompatibility entry point for failed, stalled, looping, or over-recovered Paperclip work. Use fix-failed-paperclip-run for the full diagnosis, regression, repair, real verification, and green-PR workflow; honor an explicitly requested diagnosis-only scope. | Stars: 115k | Updated | Claude CodeCodex | Verified | Install |
| fix-failed-paperclip-runpaperclipai/paperclipDiagnose and fix failed, stuck, slow, or incorrectly blocked Paperclip tasks and runs. Use with a task/run URL or failure report to trace the root cause, prove a regression red then green, check second-order effects, verify the real workflow, and deliver a PR with passing checks. | Stars: 115k | Updated | Claude CodeCodex | Verified | Install |