fr-goallisted

# fr-goal One operator touchpoint from goal to reviewed PR: brainstorm context → ONE batched Q&A → spec → review → fr-plan → review → TDD implementation → review → single PR. Every review fixes all it finds; the operator's autonomy instruction (outranking skill defaults) replaces each superpowers approval pause with a review-and-fix pass — that protection front-loads into the Q&A. When blocked: stop, state what's blocked and what you tried, and ask — a wrong guess shipped in a PR costs more than a paused run. **Announce at start:** "I'm using fr-goal to run this goal autonomously." ## What stays interactive | Moment | Why | |---|---| | The batched Q&A (once, at brainstorm) | Design decisions are operator-owned | | Multi-project repo location | Access question — only if the filesystem can't answer it | | Manual phases | Secrets, UI ops, deploys are human-only — see placement policy | | PR merge | The agent never self-merges | | Post-merge Test Plan | Proves real deployment; needs the operator's environment | ### 1. Brainstorm with batched Q&A — in isolation Invoke `fr-brainstorming` (runs `fr isolation up` first; no devcontainer profile → pause for the fr-init interview). Isolation precedes EVERYTHING — read-only exploration, measurements, and cluster ops included; an operator "start with X" changes the first work item, never the first action; from the start ALL commands run via `fr isolation exec`. Explore the workspace, collect EVERY operator-owned decision, and ask al