AI Monk Index plugin
Portable, free, read-only plugin v0.1.4 for the AI Monk Database. Root plugin.json and mcp.json follow the Agent Plugins 1.0.0 schemas. Research profiles, video structures, products and creator-published production patterns through six public tools:
| Tool | Research available |
|---|---|
search_ai_monks | Filter profiles and candidates by language, platform and other indexed fields. |
get_ai_monk | Retrieve one full profile and its referenced sources. |
ai_monk_index_stats | Inspect snapshot coverage, AI status and methodology. |
top_formats | Compare sampled video formats, hooks and script structures with dated view observations. |
how_they_monetize | Inspect products, displayed prices, seller counters and separately attributed revenue claims. |
winning_playbooks | Study source-linked script formulas, voice, cadence and product funnels published by creators. |
Candidate inclusion does not confirm AI production or video format. Sample views are not channel totals; prices and seller counters are not orders; creator claims are unaudited and do not establish profit or MRR. Unknowns remain unknown. Use playbooks as patterns rather than copying scripts verbatim. No credentials, hooks, executable, local MCP server, generation, social connection, posting, purchase flow or income guarantee is bundled.
Install in Codex
Use a Codex CLI with codex plugin marketplace support (checked with 0.160.0):
codex plugin marketplace add Daisuke134/anicca-products --sparse .agents/plugins --sparse packages/ai-monk-plugin
codex plugin add ai-monk-index@anicca
codex plugin list --marketplace anicca
This installs from Anicca's repository marketplace. The plugin connects to the existing public, read-only MCP endpoint; no API key or payment is required. Tool reference: AI Monk Index.
Demonstration
Watch the 172-second execution-log replay. It shows actual prompts, MCP results and Japanese final answers; it is not a screen recording.
The demonstration uses eight v0.1.2 runs across research snapshots v0.4.6 and v0.4.7. Six fully met the expected behavior, one recovered from an out-of-schema limit argument, and one only partially explained the public-research scope after refusing private information. These are separate runs with mixed snapshots. v0.1.3 added the verified demo URL through metadata only. v0.1.4 clarifies expected behavior for private credentials and earnings guarantees and aligns it with the MCP initialization instructions. No v0.1.3 or v0.1.4 native rerun or formal review approval is claimed.
Portable ZIP and directory submission
Use the versioned ZIP as an authoring/submission package. v0.1.2 introduced the metadata and five positive/three negative review cases for the existing six-tool endpoint. v0.1.3 adds the verified demonstration URL; v0.1.4 clarifies privacy and earnings-boundary expectations without changing the tool surface. A downloaded ZIP or reachable MCP endpoint does not establish a ChatGPT/Codex public directory listing. Initial MCP review also needs a real reviewer-accessible walkthrough and publisher authentication, identity, permission and domain verification; these cannot be fabricated in source metadata.
Release packaging uses the merged immutable main commit, with plugin.json, mcp.json, README.md and assets/icon.png at the ZIP root. Keep v0.1.1, v0.1.2, and v0.1.3 intact and publish a new ai-monk-plugin-v0.1.4 tag with ai-monk-index-plugin-0.1.4.zip. Verify its file hashes against that main commit and its public download against the official release digest. Release and directory submission remain separate operations.
The packaged icon is the existing Anicca app artwork. Code and manifest are MIT; the remote dataset has its own attribution and third-party content limits. See the service terms and privacy notice.