pmm-sherpalisted

# PMM Sherpa Calling Guide ## What Sherpa is PMM Sherpa is the advisory layer. It produces senior PMM judgment grounded in a curated corpus, in a calibrated voice (Layer 4: discovery cadence, story-first, framework-named-mid, single-question close). You are the orchestrator. Sherpa is the judgment. The four tools' own descriptions tell you *what* each tool does. This skill is about *strategy*: when to reach for Sherpa vs. handle it yourself, how to phase Sherpa across a Deep Research run, and how to keep voice consistent. ## When NOT to call Sherpa - File parsing, web fetch, web search, image generation, code drafts. Claude's native tools handle these. - Reading project knowledge or project files. Claude has direct access. - Trivial follow-ups the prior Sherpa turn already answered. - Execution tweaks ("shorter", "tighter", "swap this word"). Refine in place. - Non-PMM questions. Sherpa is a domain advisor, not a generalist. ## In regular chat Use the tool descriptions as your guide for which of the four to pick. The pattern that works: 1. **Claude gathers**: fetch URLs, parse files, search the web, pull project context. 2. **Claude summarizes the relevant constraints**: brand guidelines, ICP from prior turn, what's been tried. 3. **Claude calls Sherpa** with that context bundled into the message. 4. **Claude integrates**: present the response as the response, layer in additional analysis if it adds value. Pass project context into Sherpa's `customSystemPromptSuffix