performance-review · git:20261013.28b896d · 2026-10-13 · sha256 ae21b94e5c798859
performance-review git:20261013.28b896dA
Immutable. This exact content never changes and is served at /api/v1/blob/ae21b94e5c798859.
No tool to install first
claude plugin marketplace add https://markdownregistry.com/mp/a/art_aioovewdl4hz7fhv/ae21b94e5c7988598f891035898f2d29f6272db66a9b02bc06c7b74ca61244fc.json
Then /plugin install performance-review@markdownregistry-aioovewdl4hz7fhv in Claude Code. It installs this exact version, the one graded A here, pinned to the commit it was graded at, and leaves it there.
--- name: performance-review description: "Weekly or monthly review of your own posts: read real insights, compare to each channel's baseline, name what worked, propose next week." --- # Performance review Use this for "how did my posts do this week", a monthly report, or before planning the next batch. ## Procedure If a call answers `workspace_required`, the account has more than one workspace: show the names it lists, ask which one, and pass that `workspaceId` on every later call in this session. 1. Window: ask for the period if the person did not say (this week, last 30 days). 2. Collect posts: `list_posts` with `since` and `until` for posts made through Uplika, and `list_channel_posts` for posts written in the channels' own apps when the person wants those too. 3. For each post, `get_insights`: views, likes, replies, reposts, quotes, shares. **Null is not zero**: some channels do not report views or shares, or not yet. Keep nulls out of averages and say which channel did not report. 4. Baseline per channel: compare each post to the median of the same channel's posts in the window, never across channels (a Threads post and a YouTube video do not share a scale). With too few posts on a channel, say the baseline is weak. 5. What worked: look at the top and bottom posts per channel and name concrete differences you can see in the posts themselves: the hook type (question, number, claim, story), length, chain or single post, media or text, posting time. Say how confident you are; one post is an anecdote. 6. Say what Uplika does not have: impressions per follower, click-through rate, watch time and retention curves are not in these tools (for YouTube retention, see `yt-retention`). ## Output One table (post, channel, date, the metrics, against baseline), then three bullets for what worked, then three concrete things to try next week, each tied to a post in the table. Offer to draft next week with `schedule-week`. ## Tools - `list_posts`: posts made through Uplika in the window - `list_channel_posts`: posts on the channel, including ones written in the app - `get_insights`: per-post metrics (null means not reported)