← ClaudeAtlas

scout-rivalslisted

Use when the owner wants to know how Muse compares to the open-source agent landscape RIGHT NOW — what openclaw, hermes-agent, and other notable open-source agents shipped recently, what Muse lacks, or when the parity reservoir feels stale; also as the per-fire entrypoint of a rival-watch loop. Produces judged intelligence only — building belongs to grow-muse/improve-muse.
wlsdks/muse-agent · ★ 5 · AI & Automation · score 77
Install: claude install-skill wlsdks/muse-agent
# scout-rivals — date-anchored rival intelligence One invocation = one **delta scout**: what changed in the rival landscape since the last scout, judged through Muse's identity lens, delivered as ledger fuel for the building skills. This skill NEVER builds product code — its ship is the intelligence itself. The exhaustive base already exists: the 2026-06-23 teardown (`docs/goals/competitor-teardown.md`, 420 files → 231 judged opportunities in `growth-backlog.md` + `judgment-lens.md`). Re-deriving that ground is the documented waste this skill exists to prevent — scout the DELTA, never the base. ## The cycle 1. **ANCHOR** — run `date` (never assume); read the watermark block at the top of `docs/goals/rival-watch.md` (last scout date, per-repo upstream SHAs, roster + reference shelf). `git -C /Users/jinan/ai/<name> fetch origin` EVERY roster and shelf clone now — the fire reads today's upstream, never a stale checkout. Everything below is scoped to changes AFTER the watermark. 2. **SWEEP (delta only, verify-in-code)** — - **Named rivals** (roster in rival-watch.md carries repo URLs, SHAs, and the persistent local clones under `/Users/jinan/ai/<name>` — `git -C` fetch them, never re-clone; a NEW roster member gets a blobless clone there). High-velocity repos ship thousands of commits per window — sweep RELEASES + CHANGELOG first (observed fire 1: 7.8k commits in 3.5 weeks; raw log is for locating an implementing file, never th