Skip to content
Back to search

content-ideas

Track competitors across X, Instagram, TikTok, and YouTube, see what they post, what performs, and get content ideas backed by real engagement data. Cross-host plugin for Claude Code & Codex.

Safety scan

A static text scan, not a guarantee. Select the badge to see what was scanned and every finding.

Checked today
Badge for your README
findsafeskills safety rating

Markdown

[![findsafeskills safety rating](/api/listings/bradautomates-content-ideas/badge.svg)](/listings/bradautomates-content-ideas)

HTML

<a href="/listings/bradautomates-content-ideas"><img src="/api/listings/bradautomates-content-ideas/badge.svg" alt="findsafeskills safety rating" /></a>

Install

i

Install instructions from the README

As written by the repo's author. findsafeskills has not verified these steps.

| Surface | Install | |---------|---------| | **Claude Code** | `/plugin marketplace add bradautomates/content-ideas` then `/plugin install content-ideas@content-ideas` | | **Codex, Cursor, Copilot, Gemini CLI, +50 more** | `npx skills add bradautomates/content-ideas -g` | | **claude.ai** (web) | [Download `content-ideas.skill`](https://github.com/bradautomates/content-ideas/releases/latest) → Settings → Capabilities → Skills → `+` | | **Manual / dev** | `git clone https://github.com/bradautomates/content-ideas.git ~/.claude/skills/content-ideas` |

Claude Code

/plugin marketplace add bradautomates/content-ideas
/plugin install content-ideas@content-ideas

Update later with `/plugin update content-ideas@content-ideas`.

claude.ai (web)

1. [Download `content-ideas.skill`](https://github.com/bradautomates/content-ideas/releases/latest) from the latest release. 2. Go to Settings → Capabilities → Skills. 3. Click `+` and drop the file in.

Enable "Code execution and file creation" under Capabilities first — the skill shells out to Python, so it won't run without it.

Codex, Cursor, Copilot, Gemini CLI, and 50+ other hosts

The [Agent Skills](https://github.com/vercel-labs/skills) CLI installs the skill into whatever agents it detects:

npx skills add bradautomates/content-ideas -g

`-g` installs globally for your user (`~/.codex/skills`, `~/.cursor/skills`, etc.); drop it to install into the current project instead. Useful flags:

- `-a, --agent <names…>` — target specific hosts, e.g. `-a codex -a cursor` - `-l, --list` — list the skills in this repo without installing - `--copy` — copy files instead of symlinking (for filesystems without symlink support)

The CLI discovers the skill from `SKILL.md` and honors `.codex-plugin/plugin.json` (`"skills": "./skills/"`) and `AGENTS.md`; `SKILL.md` resolves its own scripts from the host cache or the clone.

What was scanned

The files the safety scan read, as written by the repository's author and shown as plain text. Lines with a finding are marked.

.claude-plugin/marketplace.json · Manifest · 21 lines

View this file on GitHub (read at the repository's current version: the scan did not record which commit it read)

  1. {
  2. "name": "content-ideas",
  3. "owner": {
  4. "name": "Bradley Bonanno"
  5. },
  6. "metadata": {
  7. "description": "Marketplace hosting the content-ideas plugin — competitor tracking and a curated For You feed for content creators."
  8. },
  9. "plugins": [
  10. {
  11. "name": "content-ideas",
  12. "description": "Track competitors across X, Instagram, TikTok, and YouTube. See what they're posting, what's performing, and get actionable content ideas backed by real engagement data. Ships the content-ideas skill.",
  13. "version": "2.2.0",
  14. "author": {
  15. "name": "Bradley Bonanno"
  16. },
  17. "source": "./",
  18. "category": "productivity"
  19. }
  20. ]
  21. }
README.md · README · 229 lines · 1 finding marked

View this file on GitHub (read at the repository's current version: the scan did not record which commit it read)

  1. # /content-ideas
  2. ![content-ideas hero](hero.png)
  3. **An AI content idea generator that does your morning research for you — it reads what's working across X, Instagram, TikTok, and YouTube, then hands you a daily "For You" feed of content ideas framed against your own brand.**
  4. I built this for my own YouTube channel and used it to grow from 700 to over 10,000 subscribers in three months. It's not just for YouTubers — it's for anyone who has to keep producing content: creators, businesses and marketers, personal brands, founders, agencies, newsletter and podcast writers. If your week starts with "what should I make next?", this skill is for you.
  5. Claude Code (recommended — auto-updates via marketplace):
  6. ```
  7. /plugin marketplace add bradautomates/content-ideas
  8. /plugin install content-ideas@content-ideas
  9. ```
  10. Codex, Cursor, Copilot, Gemini CLI, or any of 50+ [Agent Skills](https://github.com/vercel-labs/skills) hosts:
  11. ```bash
  12. npx skills add bradautomates/content-ideas -g
  13. ```
  14. (`-g` installs globally for your user, available across all projects. Drop it to scope per-project.)
  15. More install options (claude.ai web, manual) in the [Install](#install) section below.
  16. One [ScrapeCreators](https://scrapecreators.com) API key covers all four platforms — X, Instagram, TikTok, and YouTube (including transcripts). 100 free calls, no card. Nothing else to install: the runtime is pure Python stdlib.
  17. ---
  18. ## The problem it solves
  19. The hard part about content is everything *before* the idea: opening four apps, scrolling until something sparks, watching a pile of other people's videos, and trying to hold the patterns in your head just to figure out what's worth making this week. That research grind eats hours.
  20. `/content-ideas` takes over that one job: **idea aggregation.** It sorts through what everyone else in your niche is doing, surfaces what's actually working, and hands you a short list of ideas so you can skip the scroll and go straight to deciding what to make.
  21. It deliberately stops there. It doesn't write your scripts, it doesn't replace your creative judgment, and the ideas it gives you are exactly that — *starting points*. The best version of any idea still has to be yours: your angle, your spin, your format. Think of it as the researcher and editor handing you a briefing, not the creator. It collapses the hours of aggregation; what you build from the brief is still you.
  22. ```
  23. /content-ideas
  24. /content-ideas ai video tools
  25. ```
  26. ## Why this exists
  27. I run a [YouTube channel](https://www.youtube.com/@bradbonanno) and a business, [Solaris Automation](https://www.solarisautomation.io/), so I live downstream of the same question every day: *what should I make next?* I usually start by looking at what's working right now but doing that by hand means doomscrolling and taking notes to spot the patterns.
  28. So I built a skill to do the morning research for me, and used it on my own channel to go **from 700 to over 10,000 subscribers in three months.** It pulls what the creators I watch actually posted, ranks it by engagement instead of recency, reads the comments to see what audiences are actually asking for, and hands me the overperformers plus a few concrete starting-point ideas framed against my own pillars and voice. It's the difference between *browsing* for inspiration and *being briefed* on it.
  29. The personalization is the point. Before it recommends anything, it studies your own content to learn your niche, pillars, and voice. It then checks every idea against what you've already published so it doesn't suggest reruns. Upvote and downvote anything in the feed and it remembers using auto memory: a self-improving loop where every run sharpens the next.
  30. ## What people actually use it for
  31. **The daily brief.** `/content-ideas` first thing in the morning. One page: every tracked creator's last week of posts merged with discovered niche outliers, sorted by what's overperforming, plus up to 10 starting-point ideas.
  32. **Niche-scoped ideation.** `/content-ideas ai video tools` biases the feed and the ideas toward a topic. Useful when you're planning a series, a campaign, or chasing a specific trend rather than scanning your whole world.
  33. **Reading the comments at scale.** The skill pulls the comments off competitors' posts and reads them, so you see what audiences are confused about, asking for, or arguing over.
  34. **Competitor teardown on demand.** Drop in a handful of post URLs and get them fetched, scored, and analyzed — hook, format, comments, why it landed — without adding the account to your tracked list. Good for "how did *this* specific post blow up?"
  35. **Profiling your own presence.** During setup it studies your own channels over the last 90 days to characterize your niche, pillars, and voice, so every idea afterward is framed as *you*, not a generic trend report.
  36. ## What makes it different
  37. - **It knows your brand.** It watches your own videos and reads your posts first, learning your niche, pillars, and voice and every idea is framed as *you*, not a generic trend report.
  38. - **It reads the comments, not just the metrics.** It pulls all the comments off competitors' posts and analyzes them for the questions, requests, and arguments that point straight at your next angle.
  39. - **It gets smarter every run.** Upvote and downvote anything in the feed; reactions feed a self-improving loop that sharpens what it surfaces over time.
  40. - **It's honest about its lane.** It replaces the research-and-aggregation grind and gives you starting points, not finished scripts. The creative call stays yours.
  41. - **Four platforms today, more coming.** Reads X, Instagram, TikTok, and YouTube through one API key, with more platforms coming soon.
  42. ## How it works
  43. 1. **You run `/content-ideas`** (optionally with a topic filter). First run walks you through setup; after that it goes straight to the feed.
  44. 2. **The scraper pulls recent posts** A recency window (`--days`, default 7, capped at 90) keeps stale posts out.
  45. 3. **Everything gets scored by engagement,** normalized across platforms, with overperformers flagged against each creator's baseline and niche outliers surfaced even from accounts you don't track.
  46. 4. **Ideas are generated** against your brand profile and checked against your own content history so you don't get told to make something you already made.
  47. 5. **A single self-contained HTML page is rendered** with two tabs (Posts + Ideas), embedded data, sort/filter controls, and a reaction layer.
  48. 6. **You react** (▲ more like this / ▼ less / a note). Reactions are saved as the substrate for future personalization so the feed learns your taste over time.
  49. ### Platform notes
  50. - **X / Twitter is top-tweets, not recent.** The ScrapeCreators X endpoint has no "latest" sort — it returns an account's most *popular* tweets, which the recency window then filters down. So for X you only ever see recent tweets that are *also* high-engagement; an ordinary recent tweet won't appear. The other three platforms (Instagram, TikTok, YouTube) are fetched newest-first, so their feeds reflect everything posted in the window. Practically, expect X to be sparser and skewed toward an account's hits.
  51. - **X has no comments or transcripts.** There are no comment or transcript endpoints for X, so X posts are text-only — the comment-reading and transcript analysis only apply to Instagram, TikTok, and YouTube.
  52. ## The feed
  53. A self-contained HTML page with two tabs:
  54. | Tab | What it shows |
  55. |-----|---------------|
  56. | **Posts** | One feed merging every tracked-account post and every discovered niche outlier. Sort by Popular or Recent; filter by creator, platform, and outliers; overperformers carry an intensity-scaled badge so winners are easy to spot. |
  57. | **Ideas** | Up to 10 actionable, differentiated briefs — angle, hook, funnel + CTA, repurposing — backed by competitor performance and your own content history. |
  58. React to any item and the reaction is saved to `$CONTENT_HOME/research/{date}/feedback.json`, feeding future personalization.
  59. ## Install
  60. | Surface | Install |
  61. |---------|---------|
  62. | **Claude Code** | `/plugin marketplace add bradautomates/content-ideas` then `/plugin install content-ideas@content-ideas` |
  63. | **Codex, Cursor, Copilot, Gemini CLI, +50 more** | `npx skills add bradautomates/content-ideas -g` |
  64. | **claude.ai** (web) | [Download `content-ideas.skill`](https://github.com/bradautomates/content-ideas/releases/latest) → Settings → Capabilities → Skills → `+` |
  65. | **Manual / dev** | `git clone https://github.com/bradautomates/content-ideas.git ~/.claude/skills/content-ideas` |
  66. ### Claude Code
  67. ```
  68. /plugin marketplace add bradautomates/content-ideas
  69. /plugin install content-ideas@content-ideas
  70. ```
  71. Update later with `/plugin update content-ideas@content-ideas`.
  72. ### claude.ai (web)
  73. 1. [Download `content-ideas.skill`](https://github.com/bradautomates/content-ideas/releases/latest) from the latest release.
  74. 2. Go to Settings → Capabilities → Skills.
  75. 3. Click `+` and drop the file in.
  76. Enable "Code execution and file creation" under Capabilities first — the skill shells out to Python, so it won't run without it.
  77. ### Codex, Cursor, Copilot, Gemini CLI, and 50+ other hosts
  78. The [Agent Skills](https://github.com/vercel-labs/skills) CLI installs the skill into whatever agents it detects:
  79. ```bash
  80. npx skills add bradautomates/content-ideas -g
  81. ```
  82. `-g` installs globally for your user (`~/.codex/skills`, `~/.cursor/skills`, etc.); drop it to install into the current project instead. Useful flags:
  83. - `-a, --agent <names…>` — target specific hosts, e.g. `-a codex -a cursor`
  84. - `-l, --list` — list the skills in this repo without installing
  85. - `--copy` — copy files instead of symlinking (for filesystems without symlink support)
  86. The CLI discovers the skill from `SKILL.md` and honors `.codex-plugin/plugin.json` (`"skills": "./skills/"`) and `AGENTS.md`; `SKILL.md` resolves its own scripts from the host cache or the clone.
  87. ## First run
  88. The first `/content-ideas` walks you through setup — no separate install step:
  89. 1. **API key.** Grab a free [ScrapeCreators](https://scrapecreators.com) key (100 calls, no card) and the skill stores it in `~/.config/content/.env` (one key covers all four platforms).
    • Dismissed after review: Tells the agent to read credentials and send them somewhere or keep it from the user
    • Line 139 on GitHub
  90. 2. **Your profile.** Share your own handles and the skill scrapes them over a full quarter to characterize your niche, pillars, and voice — the basis for every idea afterward.
  91. 3. **Tracked competitors.** Add the creators you want to watch, per platform.
  92. After that, `/content-ideas` goes straight to the feed. A SessionStart hook prints a one-line nudge only when a key is missing — silent once you're set up.
  93. ## Usage
  94. ```
  95. /content-ideas # your full daily feed
  96. /content-ideas ai video tools # bias the feed + ideas toward a topic
  97. ```
  98. Driving the scraper directly (from a checkout):
  99. **Profile mode** — recent posts for tracked accounts:
  100. ```bash
  101. python3 skills/content-ideas/scripts/scrape.py '{"x": ["h1"], "instagram": ["h2"]}' \
  102. --pillars "..." --since 2026-04-15 --days 7
  103. ```
  104. `--days N` keeps the last N days (default 7, hard-capped at 90); `--since` can only narrow that window further. The small default keeps the daily feed tight; the wide end is for the one-off profile-building scrape during setup.
  105. **URL mode** — fetch individual posts for ad-hoc analysis:
  106. ```bash
  107. python3 skills/content-ideas/scripts/scrape.py urls \
  108. "https://x.com/u/status/1" "https://instagram.com/p/abc" --pillars "..."
  109. ```
  110. Both modes require a ScrapeCreators key. With none configured the scraper stops and shows setup instructions — there is no silent fallback to web search.
  111. ## Where your data lives
  112. Everything persists under **`$CONTENT_HOME`** (the `CONTENT_HOME` env var, or `~/Documents/Content` by default) — one stable home, independent of where you invoke the skill:
  113. ```
  114. $CONTENT_HOME/ (default ~/Documents/Content)
  115. ├── brand/ Context (built during first-run setup)
  116. │ ├── profile.md Niche, audience, platforms, pillars, search terms, goal
  117. │ ├── my-content.md Own content performance + audience requests
  118. │ └── tracked-accounts/{platform}.md Tracked creators per platform
  119. └── research/{YYYY-MM-DD}/ Output, one dated folder per run
  120. ├── feed-data.json Feed data (skill-generated)
  121. ├── for-you.html Rendered feed (generate_feed.py --static)
  122. └── feedback.json Your reactions (when you react in the feed)
  123. ```
  124. Credentials live separately in `~/.config/content/.env`. See [FILE-SCHEMAS.md](FILE-SCHEMAS.md) for the exact structure of every file.
  125. ## Structure
  126. ```
  127. .
  128. ├── skills/content-ideas/
  129. │ ├── SKILL.md Skill contract — loaded by all surfaces
  130. │ ├── references/content-strategy.md Angle/hook/funnel/CTA/brief domain knowledge
  131. │ ├── assets/for-you-template.html For You page template (data embedded by the generator)
  132. │ └── scripts/
  133. │ ├── scrape.py Scraper entrypoint (profile / urls / setup)
  134. │ ├── generate_feed.py Embeds feed data into the template; serves or writes static HTML
  135. │ └── lib/ Small, unit-tested modules (http, platforms, scoring, relevance, ...)
  136. ├── commands/content-ideas.md Claude Code slash command
  137. ├── hooks/ SessionStart setup-status hook
  138. ├── .claude-plugin/ plugin.json + marketplace.json (Claude Code)
  139. ├── .codex-plugin/ plugin.json (Codex)
  140. ├── AGENTS.md → CLAUDE.md Codex / generic-agent entry point
  141. ├── scripts/build-skill.sh Builds dist/content-ideas.skill for claude.ai upload
  142. ├── tests/ Unit tests (python3 -m unittest discover -s tests)
  143. └── .github/workflows/ validate.yml (tests) + release.yml (auto-builds the .skill on tag)
  144. ```
  145. ## Develop
  146. ```bash
  147. # Run the test suite (stdlib only — no pip install needed):
  148. python3 -m unittest discover -s tests -p "test_*.py"
  149. # Build the claude.ai upload bundle:
  150. bash scripts/build-skill.sh # → dist/content-ideas.skill
  151. ```
  152. Releasing: tag `vX.Y.Z`, push the tag. The workflow builds `dist/content-ideas.skill` and attaches it to the GitHub release. Keep the version in sync across `pyproject.toml`, `.claude-plugin/plugin.json`, `.claude-plugin/marketplace.json`, `.codex-plugin/plugin.json`, and `SKILL.md` — `tests/test_plugin_contract.py` enforces it.
  153. ## Open source
  154. MIT license. Runtime is dependency-free Python stdlib; scraping via the [ScrapeCreators](https://scrapecreators.com) API.
  155. Built by Brad Bonanno — I make content about building with AI on [YouTube (@bradbonanno)](https://www.youtube.com/@bradbonanno), and build AI operating systems for businesses at [Solaris Automation](https://www.solarisautomation.io/). If this skill saves you a morning scroll, come say hi on the channel.
  156. ---
  157. [github.com/bradautomates/content-ideas](https://github.com/bradautomates/content-ideas) · [@bradbonanno](https://www.youtube.com/@bradbonanno) · [Solaris Automation](https://www.solarisautomation.io/) · [LICENSE](LICENSE)

Ratings and comments

Sign in to rate this skill.

Sign in

Free to use. If it’s helped, pass it on.