gemini-managed-agents-templates

Mix speech audio and background music into a polished radio show file.

Generate cover images for the radio show using Gemini 3 Flash Image (with fallback to Pro).

Generate show metadata (JSON) containing timecoded transcript, title, duration, summary, and date by sending audio and transcript to Gemini.

Generate ambient background music using Lyria via the Interactions API.

Gather source material for the radio show from any content source — web, HN, GitHub, papers, or general topics.

Generate the AI Talk Radio show script from research using the Interactions API.

Convert radio show script to speech audio with telephone effect on correspondent voices using the Interactions API.

Records and appends every user-agent customer support interaction to the memory.md log in the workspace.

Scans a website deeply, converting HTML pages to markdown, respecting robots.txt, and updating the snapshots log.

General-purpose data profiling and exploration. Use when first encountering any dataset to understand its structure, quality, and analysis potential.

Performs heavy numerical compute and statistical analysis inside the sandbox.

Exposes a 100% local, offline PDF batch extraction utility (extract_to_markdown.py) that isolates invoices under invoices/ and translates PDFs into clean Markdown files for LLM-native parsing.

Reconcile loaded expenses against the pre-parsed invoice database, flagging discrepancies like amount mismatches, missing invoices, and merchant mismatches locally.

Create Google-style presentation slides as standalone HTML files. Use when the user wants to (1) create presentation slides from notes or bullet points, (2) build slide decks in HTML format, (3) recreate or replicate existing slide designs, (4) convert rough notes or talking points into polished visual slides, or (5) needs professional Google-branded presentation materials with clean typography, visual diagrams, code blocks, or data visualizations.

Verify whether vendors/merchants from expense records are real, registered business entities by querying the free Wikidata open API.

Formats the standalone daily briefing — curated picks with why-it-matters lines, community buzz, the trend read across runs, and the taste note that keeps the learning visible.

Selection and ranking rules — how to turn the raw fetched pool into the handful of items the reader actually wants, guided by the taste profile in interests.json.

Posts a condensed version of the briefing to a Google Chat or Slack incoming webhook, so the daily run delivers itself — skips silently when no webhook is configured.

Pulls the latest Google News and Hacker News items for every topic and genre in the reader's interests, deduped against every item shown in previous runs.

Guides the agent on how to handle git operations and generate .patch files instead of submitting PRs.

Guides the agent on how to read and search GitHub issues to understand the repository's problems.