From glean
Review delivered glean fragments interactively - star, archive, or trash each one, capture a short note on why, feeds signals into feedback.jsonl
How this skill is triggered — by the user, by Claude, or both
Slash command
/glean:glean-reviewThis skill is limited to the following tools:
The summary Claude sees in its skill listing — used to decide when to auto-load this skill
Walk the user through their inbox of newly delivered story fragments. Capture star/archive/trash decisions plus an optional note. This feeds future auto-tuning.
Walk the user through their inbox of newly delivered story fragments. Capture star/archive/trash decisions plus an optional note. This feeds future auto-tuning.
Fetch the review queue. The worker prints a JSON array of inbox fragments (status: inbox, sorted by captured desc):
PYTHONPATH="${CLAUDE_PLUGIN_ROOT}/scripts" python3 -m glean.worker --review
Each item should include at minimum: fragment_id, path (absolute), title, project, captured, draft_angle (from the ## Thread section), and an excerpt (first paragraph of ## What happened or similar).
If the array is empty, print Inbox is empty. Nothing to review. and exit.
Initialize tallies: starred=0, archived=0, trashed=0, skipped=0. Track remaining = len(items).
For each fragment, display a compact card:
[i/N] {title}
project: {project} captured: {captured}
thread: {draft_angle}
{excerpt}
path: {path}
Ask the user with AskUserQuestion:
Star, Archive, Trash, Skip, QuitIf the action is Star or Trash, ask a follow-up via AskUserQuestion for a one-line note on WHY (especially important for trash). Allow an empty note. Skip the note prompt for Archive and Skip unless the user wants to add one.
Apply the decision:
a. Update the fragment file's frontmatter status field using Edit:
Star -> status: starredArchive -> status: archivedTrash -> status: trashedSkip -> leave as inbox, do not write feedbackb. Append a feedback entry to ${CLAUDE_PLUGIN_DATA}/feedback.jsonl (one line, JSON):
{"fragment_id": "...", "action": "star|archive|trash", "note": "...", "timestamp": "<ISO-8601 UTC>"}
Prefer calling the worker if it exposes a feedback-writing subcommand; otherwise append directly with a shell heredoc. NOTE: CONTRACTS.md does not specify a dedicated feedback CLI flag - if none exists, append via shell and say so clearly in a comment to the user.
After each decision, print the running tally and remaining count:
tally: * {starred} starred | = {archived} archived | x {trashed} trashed | - {skipped} skipped | remaining: {remaining}
If the user picks Quit, stop immediately, print a final summary, and exit. Do not process remaining items.
After the loop ends (all reviewed or user quit), print a final summary:
glean-review session summary
----------------------------
reviewed: N starred: N archived: N trashed: N skipped: N
feedback written to: ${CLAUDE_PLUGIN_DATA}/feedback.jsonl
Trash only changes the frontmatter status.Z suffix.npx claudepluginhub vedantggwp/gleanProvides UI/UX resources: 50+ styles, color palettes, font pairings, guidelines, charts for web/mobile across React, Next.js, Vue, Svelte, Tailwind, React Native, Flutter. Aids planning, building, reviewing interfaces.
Fetches up-to-date documentation from Context7 for libraries and frameworks like React, Next.js, Prisma. Use for setup questions, API references, and code examples.