Changelog

All notable changes to this tile are documented here.

Changed — per-skill agentModel: tier-down (jbaruch/nanoclaw#613)

Pin the cadence skills' models via agentModel: frontmatter so they stop defaulting to Opus: Sonnet (claude-sonnet-4-6) for entertainment-sync — it synthesizes watch/read recommendations (its Skill()-invoked recommend-* sub-skills run in the same spawn, so recommendation quality rides on this model, not Haiku); Haiku (claude-haiku-4-5-20251001) for check-watchlist and youtube-comment-check (triage). Part of the #613 Claude tier-down.

0.1.0

Added

• Initial tile: the personal entertainment-media skill cluster migrated from nanoclaw-admin into a standalone public per-chat overlay tile (jbaruch/nanoclaw-admin#296). Seven skills move together because they share intra-cluster data under /workspace/group/ (e.g. entertainment-sync and trakt-watch-history write trakt-history.json, which recommend-shows reads; recommend-shows writes watchlist.json, which check-watchlist reads; audible-backup writes books-library.csv, which recommend-books reads): entertainment-sync (weekly cadence wrapper), recommend-shows, recommend-books, trakt-watch-history, check-watchlist (nightly cadence), youtube-comment-check (weekly cadence), and audible-backup. Carries each skill's helper scripts, state-schema/reference docs, and unit tests unchanged from the admin originals. The cluster is self-contained — no cross-tile code dependency on nanoclaw-admin's heartbeat/Composio shared infra; each skill talks to its own data plane (Trakt API, YouTube Data API, owner-uploaded CSVs).

Rules

• Closed-loop carve-out claimed for jbaruch/coding-policy: plugin-evals (2026-06-07). This tile is part of the jbaruch/nanoclaw-* plugin fleet — a fully-automated agent loop satisfying all three preconditions of the rule's "Narrow exception for closed-loop automated systems with no human eval-result consumption" clause: (1) no human reviews eval output for this tile in any form (no eval scores, no lift deltas, no scenario-by-scenario diffs, no regression alerts); (2) no automated gate consumes eval results (no evals.yml workflow, no publish-tile eval step, no downstream dashboard or paging route); (3) the owner accepts that re-introducing any consumption of eval results later — whether human review OR automated gating — requires re-introducing evals first under the standard requirement. Matches the carve-out claimed by jbaruch/nanoclaw-admin on 2026-05-09 and inherited by every jbaruch/nanoclaw-* tile thereafter. Covers all seven skills in this tile. No evals/ directory ships in this tile.