routerlisted

Map of the agent-memory skills and the chains they are meant to run in. Read it when the user sets a task and the route is not obvious: unclear which skill solves it, the task has several steps, or you are about to do by hand what a skill already covers. Also on direct words: "which skill", "where do I start", "what is the best way to do this", "is there a skill for this".
letsloose501/sqs-agent-memory · ★ 1 · AI & Automation · score 77

Install: claude install-skill letsloose501/sqs-agent-memory

# Route Name the route and offer the first step; do not recite the skill list. One line in the user's language, like "route: `/agent-memory:grilling` -> `/agent-memory:to-spec` -> `/agent-memory:to-tickets` -> `/agent-memory:implement`, start with the interview?", and that is all. Do not ask permission for every next step: ask at forks. If a skill covers the task and you already started by hand, stop and offer the skill. If no skill covers it, say so and work by hand; do not stretch the nearest one over it. All skills below live in the `agent-memory` plugin: call them as `/agent-memory:<name>`. ## Forks that are easy to get wrong **Text in hand versus text to extract.** Ready text, a summary, notes: straight to `notes`. A video link: `video`, which extracts the transcript and calls `notes` itself. A PDF, a book, a scan: `pdf`, which also calls `notes` itself. Never call `notes` directly on a link or a file. **A document to learn from, or a file to work on.** A PDF whose content goes into the vault: `pdf`. A Word, PowerPoint or Excel file to create, read, edit or convert, or a PDF to merge, split, fill or sign: Anthropic's `document-skills` (`docx`, `pptx`, `xlsx`, `pdf`; in the Claude apps they show as `anthropic-skills:*`). Not installed: say so and offer stage 4b of `setup`; do not improvise with raw libraries when the user asked for a finished document. **Knowledge from a video, or how the video is made.** `video` is for knowledge: a lecture, a talk, a tutorial, a p