Disciplined software changes: understand the requirement, reproduce failures, make the smallest safe change, and verify it with evidence. Use when writing, modifying, debugging, or reviewing code, or when a task produces anything a compiler, interpreter, linter, or test suite will evaluate.
Decomposes complex research questions into independent sub-questions, researches each in its own unit, then synthesizes a cited report. Use when a question requires multiple independent angles of investigation. Do not use when a single research path is sufficient to answer the question.
Searches, reads, and aggregates Gmail content, finding bookings, costs, topics, or senders, and summarizing or digesting threads. Use when the ask involves email in any way: finding a message, extracting amounts or dates from mail, summarizing a thread, or digesting what a sender or label contains.
Source-grounded factual research with verification and citations. Use when a question requires current or externally verifiable facts and a single research path is sufficient. Use deep-research when multiple independent research angles are required.
Carries a hackathon from its published rules to a submission-ready project. Extracts every rule and constraint first, ranks ideas that all comply, and builds toward the exact deliverables the judges expect. Use when a mission is given a hackathon link or an inline hackathon brief. Do not use for general coding or…
Drafts, rewrites, and edits prose in the owner's own voice, in Bangla or English, and keeps factual content controlled instead of fluent guesswork. Use when the deliverable is text a person will read or send: an email, a message, a post, an article, a document, a summary for humans, or a rewrite of any of these. Do…
★not rated 59 yesterdayA0 tokens
AGPL-3.0
At most 3 mods per repository are shown here, and a mod shipped inside a plugin is left to that plugin's page — the rest are on their repository pages: