Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx skills add chuwu57/Chuwu-PR-OS --skill spoken-content-extractorgit clone --depth 1 https://github.com/chuwu57/Chuwu-PR-OSWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/skills/chuwu57/chuwu-pr-os/spoken-content-extractor)<a href="https://agentmods.dev/skills/chuwu57/chuwu-pr-os/spoken-content-extractor"><img src="https://agentmods.dev/badge/skills/chuwu57/chuwu-pr-os/spoken-content-extractor/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/skills/chuwu57/chuwu-pr-os/spoken-content-extractor"><img src="https://agentmods.dev/badge/skills/chuwu57/chuwu-pr-os/spoken-content-extractor.svg" alt="Reviewed on agentmods" width="80" height="20"></a>What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00105 | $0.01762 |
| Opus 5.5 | $0.00042 | $0.00705 |
| Sonnet 5.5 | $0.00021 | $0.00352 |
| Haiku 4.5 | $0.00011 | $0.00176 |
Grade A, and why
spoken-content-extractor scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 13d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 114 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Spoken Content Extractor
归属说明：本Skill由初五维护，属于初五的个人Skill体系。分享或再分发时请保留本归属说明及已有来源、作者和许可声明。
Core Rule
Work only from materials the user can access or has provided. Do not bypass paywalls, private rooms, login-only content, anti-scraping controls, DRM, or platform restrictions. If a link cannot be accessed directly, ask for an exported audio file, copied page text, transcript, show notes, or another accessible source.
Artifact Placement
Keep execution material and knowledge assets separate from the start:
- Working layer: downloaded audio, temporary dependencies, ASR Markdown, JSONL, logs and inspection files go under the current task's
work/. They are not active knowledge-base content. - Source preservation: only when the user asks to preserve source material for future verification, place original audio and full transcript in the appropriate
Archive/or60-References/location, with provenance and access boundary. Do not place copyrighted full text in an active knowledge entrypoint. - Active knowledge layer: when the user explicitly asks to absorb the episode into the knowledge base, write only evidence-backed facts, short verified quotations, structured Reading or a reusable framework/case. Link to the preserved evidence record; do not copy the full ASR transcript into Domains, README or MOC.
- No automatic write-back: extraction, brief and learning-guide requests stay in task outputs unless the user explicitly asks for knowledge-base absorption.
Mode Gate
Choose the least expensive mode that can support the requested conclusion. Accessible audio is not, by itself, a reason to run full ASR.
- Knowledge-increment mode — default for「吸收、归档、看看什么值得入库、提炼长期判断」:
- Inspect public metadata, detailed show notes, timeline, linked primary material, and existing knowledge assets first.
- State the「新增回答的问题」before transcription. If the episode adds no mechanism, evidence, state, boundary, or action, link it as supporting evidence or skip write-back.
- Use show notes and targeted verification where sufficient. Mark the Evidence
partialand disclose coverage when part of the episode has not been transcribed or heard.
- Targeted-verification mode — default when one or more pivotal claims need audio support:
- Use the timeline to transcribe only the relevant windows, normally 2–5 minutes around each target.
- Verify exact wording against the audio before quoting. Do not upgrade a targeted sample into a whole-episode conclusion.
- Full-transcript mode — use only when:
- the user explicitly asks for a complete transcript, complete episode summary, or full-coverage study guide;
- exact quotation coverage across the whole episode is required; or
- usable show notes/transcripts are absent and the requested conclusion genuinely requires whole-episode coverage.
What ships with it
6 files beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 13d ago Changed · -79 lines · +10 tokens per session 861048d938b6
- 1mo ago First seen · 193 lines · 95 tokens per session scan A 90a588c9a8b0
spoken-content-extractor is a skill published in the GitHub repository chuwu57/Chuwu-PR-OS (84 stars, last pushed 13d ago), licensed MIT. It adds 105 tokens to every session and 1,762 once invoked, about $0.0004 per session on Opus 5.5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-09-08.
Other skills, from other repositories
goal
Use when the user invokes /goal, asks for goal mode, designs an agent loop, or starts long-horizon autonomous work that should be clarified into a concrete, verifiable objective before execution.
launch-loop-strategy
Use when planning a practical launch loop for a feature, product update, template, prompt, skill, integration, beta, or release with a known launch date, especially when the user needs channels, content angles, a posting schedule, community distribution, replies, DMs, and follow-up remixes.
marketing-growth
Use when planning product marketing, writing marketing copy, designing pricing tiers, building programmatic SEO pages, optimizing on-page SEO, managing social media accounts and publishing queues, planning social media calendars, running Twitter automation, conducting Google Trends research, brainstorming brand/domain…
launch-strategy
Use when planning a product launch, feature release, waitlist, beta, public announcement, relaunch, Product Hunt-style push, repeated feature launch system, launch-week checklist, or multi-channel GTM rollout.
x-article-writer
Use when writing or rewriting longform X/Twitter articles, X Articles, founder essays, launch essays, growth teardowns, product breakdowns, build-in-public stories, or sharp startup takes that need X-native hooks, tone, pacing, and structure. Use for X Articles and long posts, not blog SEO posts, lifecycle emails, or…
copywriting
Use when writing or rewriting website copy: homepages, landing pages, pricing pages, product pages, feature pages, hero sections, headlines, subheads, CTAs, value propositions, taglines, and above-the-fold messaging. Do not use for email sequences, blog posts, longform X articles, or platform-specific ad copy.