AdvertiseNewsletter

Daily reports › 2026-11-13 · Is your tool safe to install today?

Daily install report: 2026-11-13

On 2026-11-13 we ran the real install of 427 agent add-ons across Claude Code, Codex, Grok Build, Cursor and Pi in a clean sandbox: 863 of 1070 installs passed, 207 failed (25 of those only need an API key). Compared with the previous run (2026-11-12, 866/1070): 0 confirmed broke, 0 really fixed, 5 flaky.

New test baseline since 2026-11-12: newer releases of all 5 agents (Claude Code 2.1.319, Codex 0.167.0, Grok Build 1.0.77, Cursor 2026.11.05, Pi 1.5.0). Scores before 2026-11-12 came from older agent versions and are not like-for-like. "Confirmed broke" and "really fixed" only compare nights on the same baseline. What changed

Our mistake: the new sandbox image was also missing uvx, the tool that starts Python MCP servers. So MCP time, MCP fetch, MCP git and claimidx (20 installs across all 5 agents) failed on Nov 12 and 13 for our reason, not theirs. All 20 passed on Nov 8. Treat those 4 as untested until the next run on the fixed image. The same Nov 12–13 runs also charged a first-time npm download against each agent's MCP connect timeout (Claude Code 30s, Cursor 60s), so several working npx servers failed for our reason. The harness now pre-fetches those packages into a shared npm cache before the health check; connect timeouts themselves are unchanged and still each agent's default.

Get the weekly scorecard by email

Daily reports live on the site only, we never email them. The free weekly email (Tuesdays) has the week's install scorecard plus new tools.

Loading the signup form… or subscribe on Beehiiv

Form not loading? Subscribe on Beehiiv →

Tonight's verdicts by agent

AgentInstallsPassedNeeds an API keyBroken (2+ nights)Pending re-checkFlaky
Claude Code37133253222
Codex1198652800
Grok Build38634553332
Cursor542352600
Pi1407755801

How to read this: Needs an API key means the add-on asked for a key or token our keyless sandbox doesn't have, so it's probably fine on your machine once you add yours. Pending failed tonight (after a retry) but worked last night; we only call it broken if it fails again tomorrow. Flaky flip-flopped lately or only passed on retry.

Standout pick

MCP sequential-thinking (reference): installed cleanly in 5 agents (Claude Code, Codex, Grok Build, Cursor, Pi), 92k GitHub stars. Dynamic, reflective problem-solving through thought sequences

How we pick: clean in every agent we tested it in, nothing flaky or pending, new and really-fixed add-ons first, then GitHub stars. Never sponsored.

Confirmed broke (0)

Nothing newly confirmed broken (failing two nights in a row after a good night).

Really fixed (0)

Nothing came back after failing two or more nights.

Pending re-check (5)

Failed tonight, worked last night. Not calling these broken yet.

Flaky (5)

These flip-flopped between pass and fail recently. They pass tonight, but give them a retry if one hiccups on you.

Needs your API key (25)

Not broken, just shy without credentials. Our sandbox has no keys on purpose.

Everything failing 2+ nights (177)

Agent versions tested

Run started 2026-11-13T02:00-05:00. Every failed install is retried once in a fresh sandbox before it counts. Each skill page has the log. How we test · Weekly scorecards · All daily reports

SkillMaxx Weekly

The week's install scorecard across all 5 agents, plus new skills, plugins and MCP servers and whatever changed in the official docs. Every Tuesday. (The daily report stays on the site, no daily emails.)

Loading the signup form… or subscribe on Beehiiv

Form not loading? Subscribe on Beehiiv →

↑↓ to move · Enter to open · Esc to close