Spot Bilibili Low-Follower Viral Videos for Growth

Where this fits

A creator wants to launch a Bilibili channel around AI tools but doesn't know which topics can break through without an existing audience. Manually sorting through search results and comparing each UP主's follower count with views is tedious. This skill turns that task into a single agent command: it finds Bilibili videos that significantly outperform the creator's follower base and returns them as ranked topic references.

Why agents benefit

• Translates a broad niche into focused Chinese and product-name keywords, then runs Bilibili searches without needing an API key.
• Ranks videos by 播放粉丝比 and engagement signals (点赞, 投币, 收藏, 分享, 评论, 弹幕), not only raw view counts.
• Filters by max followers and minimum plays, so the agent can surface breakout videos from small accounts instead of established influencers.
• Outputs structured markdown that can be fed directly into a topic list, spreadsheet, or content calendar.
• Uses conservative rate limits, retries, and a local follower cache, making repeated agent runs safer and faster.

Practical scenarios

Niche research for a new channel

The user asks for "B站低粉爆款" in AI painting. The agent converts the query into keywords like "AI绘画,Stable Diffusion,Midjourney", searches recent uploads, filters to accounts under 50,000 followers, and returns videos with high play-to-follower ratios. The user can then identify recurring themes worth producing.

Validating a content hypothesis

Before spending time on a video, the agent compares multiple keyword sets — for example "AI编程" versus "大模型工具". If one topic consistently produces low-follower viral hits over the last 30 days, that is a stronger signal for topic selection than a single trending video.

Competitive reference for small creators

A small UP主 wants to study the packaging of successful peers. The agent compiles videos by low-follower creators with high engagement, showing title phrasing, cover style patterns, and comment topics that may have contributed to the viral result. This gives practical references for improving retention and click-through.

Add it to your agent workflow

Install the skill with:

npx skills add kangarooking/kangarooking-skills --skill bilibili-viral-topic

Then ask your agent to generate keyword ideas or run the bundled script directly:

python3 scripts/search_bilibili_viral_topic.py --topic "AI工具,大模型,AI编程" --days 30 --max-followers 100000 --min-play 10000 --limit 50 --format markdown

No API key required. If public requests are rate-limited, the agent can set BILIBILI_COOKIE as an environment variable without writing it into skill files or output.

Compare before adopting

Check whether the returned rankings match the metrics your use case actually needs: an account-growth workflow depends on 播放粉丝比, while a general research task may care more about absolute play counts. Because Bilibili's endpoints can change, watch the repository's maintenance freshness. Alternatives include manual search, third-party analytics dashboards, or browser automation, but this skill is lighter and faster for agents. Also confirm it fits your current workflow: it requires Python and can run as a one-off script, not a browser plugin.

Why it is worth tracking

This skill has 569 stars and is authored by kangarooking, which signals community interest in Bilibili growth research. The value is highest when you are building content-research agents for Chinese video platforms. Re-evaluate it when the repository updates its ranking logic or if you need additional filters such as date ranges, keyword expansion, or export formats. For a continuously curated view of this and similar skills, see https://www.openagentskill.com/skills/kangarooking-bilibili-viral-topic.