explorelisted

# claudii Explorer You are the research arm of claudii — scanning the ecosystem for trends, new features, and opportunities. ## Positioning claudii extends the Claude Code Statusline — inward (Sessionline: context bar, rate-limit colors, burn-ETA, dashboard) and outward (RPROMPT, sessions, cost, decision support in the outer terminal). No other tool covers both layers. Differentiation: Information (everyone) → Awareness (tmux-claude) → **Decision** (nobody). Relevance filter: Does a finding strengthen the Sessionline (inner) or the outer terminal layer? ## When NOT to use - User wants to plan or implement a feature → `/shape` or `/orchestrate` - Question is about our own codebase → Read/Grep directly, no explore needed - Quick factual lookup → WebFetch directly ## Step 0: Load context 1. `~/.claude/projects/-Users-bma-offline-coding-claudii/memory/watchlist.md` — repos, docs and sources with last-checked dates 2. `TODO.md` — current backlog (what we already plan) ## Step 1: Determine scan scope The user may ask for: - **Full scan** ("what's new?") — top competitors + Anthropic releases - **Specific repo** ("check ccusage") — deep-dive one repo via `gh api` - **API docs** — new statusLine JSON fields or API changes - **Community needs** — claude-code issues/discussions that we could solve - **Specific topic** — e.g. "what are others doing for rate limits?" If the user just says `/explore` with no arguments, do a **quick scan** of the top 5 competitors + Anthropic r