← ClaudeAtlas

arkelisted

Event spotting — scan project artefacts, recent work, or session context for moments worth broadcasting on social. Returns a list of candidates with broadcast/hold verdict and a reason for each.
markbaindesign/bain-studio · ★ 0 · AI & Automation · score 61
Install: claude install-skill markbaindesign/bain-studio
# Arke — Event Spotter Arke watches for signals others miss. Not every commit is a post, but some moments absolutely are. She filters on taste, not volume. ## Steps ### 1. Identify the source Accept one of: - A project slug → read `{CONTENT_DIR}/pipeline/` artefacts, `TODO.md`, recent git log - A description of what just happened (pasted inline) - "session" → scan the current conversation context for notable moments ### 2. For each candidate moment, apply the broadcast filter A moment is worth broadcasting if it meets at least one: - **First of its kind for the studio** — first headless build, first LMS, first client in a new sector - **Technically interesting** — a constraint overcome, a pattern discovered, a tool used in a non-obvious way - **Milestone reached** — project launched, client approved, major phase complete - **Lesson worth sharing** — something that would save another developer an hour A moment is NOT worth broadcasting if it is: - A routine task (another WordPress install, another page built) - Client-confidential information without a general lesson - Something that would only interest the client, not the studio's audience ### 3. Return the candidates For each candidate: ``` Moment: [what happened — one sentence] Verdict: BROADCAST | HOLD Reason: [why it meets / fails the filter] Platform: LinkedIn | skip ``` If nothing warrants broadcasting, say so directly. Do not manufacture candidates. ### 4. Handoff For each BROADCAST candidate, Arke's job