mirror of
https://github.com/thedotmack/claude-mem.git
synced 2026-09-20 04:23:02 +08:00
Alex Newman
3f34a8d657
Three new mode profiles that capture the strategic layer of GStack-powered development sessions — scope decisions, demand signals, design choices, quality gates, eureka moments, and shipping events. - code--gstack: inherits from code mode, adds 6 strategic types + 7 concepts - gstack: standalone strategic-only mode (YC partner's notebook) - gstack--chill: only records pivotal moments
9 lines
1.9 KiB
JSON
9 lines
1.9 KiB
JSON
{
|
|
"name": "GStack Strategic (Chill)",
|
|
"prompts": {
|
|
"recording_focus": "WHAT TO RECORD (SELECTIVE MODE)\n--------------------------------\nOnly record strategic moments that would be painful to rediscover:\n- Scope decisions that changed the product direction\n- Demand evidence that validated or invalidated a hypothesis\n- Eureka moments — first-principles insights that reframe the problem\n- Major shipping events (launched to real users, not just merged a PR)\n- Quality gate FAILURES that blocked shipping or changed the plan\n\nSkip the incremental stuff:\n- Quality gates that just passed cleanly (record only when they find something)\n- Design choices that follow established patterns\n- Routine architectural decisions matching existing system style\n- User insights that confirm what was already known\n\n✅ GOOD (would be painful to rediscover):\n- \"Killed the dashboard — users check Slack, not dashboards. Ship insights to them.\"\n- \"CEO review expanded scope after finding 3/5 users manually do X every morning\"\n- \"QA found SSE reconnect race condition — the fix changes the reconnect architecture\"\n\n❌ SKIP (obvious from context):\n- \"Chose React for the frontend\" (duh)\n- \"QA passed, all tests green\" (expected)\n- \"Design review approved the color palette\" (incremental)",
|
|
|
|
"skip_guidance": "WHEN TO SKIP (BE VERY LIBERAL)\n-------------------------------\nSkip everything except pivotal moments:\n- All routine operations, file reads, git work\n- GStack skill runs that produce expected outcomes\n- Quality gates that pass without findings\n- Design choices following established system\n- Incremental user insights confirming known hypotheses\n- Any observation where the title would be boring\n\n**If it wouldn't make Garry sit up in his chair, skip it.**\n**When in doubt, skip it. Only record what would change someone's approach.**"
|
|
}
|
|
}
|