Fleet-wide Glama score tracker -- query per-tool TDQS grades.

Identifies worst-scoring tools, trigger rescans. When to use: check fleet repo Glama scores, find tools needing docstring fixes, or trigger a refresh. When NOT to use: for code diagnostics use the repo's own tools; for Glama submission/review use glama.ai directly. Preconditions: repo must have been submitted to Glama and scored.

Operations:

• list: All repos, worst-first.

• get: Per-tool breakdown for one repo (repo_name required).

• worst_tools: Lowest-scoring tools fleet-wide.

• refresh: Rescrape all fleet repos + snapshot.

• history: Recent refresh log entries.

• staleness: Repos >7d since last scrape.

• report: Full daily status report.

• deltas: Score changes between last two snapshots.

• add_repo: Add a repo to track (name, optional author/slug).

• remove_repo: Stop tracking a repo.

• reload_config: Reload fleet-repos.json config file.

Return compact fleet-wide score summary.

Grades per repo and count per grade bucket. Returns: success, grade_distribution, repos, count, message.

Generate daily fleet health report.

Includes grade distribution, recent changes, worst tools, stale repos. When to use: once daily for comprehensive fleet Glama score overview. When NOT to use: for single repo details use glama_status(get). Preconditions: at least one refresh must have been run.

Show fleet Glama score overview as a rich Prefab card.

Displays grade distribution, per-repo scores (worst first), worst tools across the fleet, and stale repos in a structured dashboard card suitable for in-chat viewing.

Return Format

ToolResult with PrefabApp card and plain-text fallback.

Examples

await show_glama_status_card()

Show per-repo Glama score breakdown as a rich Prefab card.

Displays overall grade, TDQS scores, coherence dimensions, maintenance grade, and per-tool scores for a single repo.

Analyze Glama scores with LLM sampling and generate fixable todos.

Uses the connected LLM (via MCP ctx.sample) to intelligently analyze a repo's tool docstring scores and produce actionable fix tasks.

Generate markdown analysis reports for repos.

Creates per-repo fix-todo reports in the reports/ directory, suitable for ingestion by IDE LLMs. If repo_name is empty, generates reports for all scored repos.