What is slowest-tests?

A tool for finding the slowest test files in a recent Buildkite run. Buildkite is a continuous-integration service that runs automated checks; the tool can also format the results for posting to Slack.

Use it to inspect a recent or specified build, produce a top-N slow-test table or JSON file, and optionally post the table to a Slack channel.

It shows which tests are making CI runs slow, so the team can investigate the biggest delays instead of guessing.

slowest-tests — Skill for Claude Code by twaldin

Install

Getting it into your agent

One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.

Any agent

npx skills add twaldin/hone --skill slowest-tests

Clone the repo

git clone --depth 1 https://github.com/twaldin/hone

Made for: Claude Code.

Wrote this? Show the measurements

A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.

README.md

[![agentmods](https://agentmods.dev/badge/skills/twaldin/hone/slowest-tests/github.svg)](https://agentmods.dev/skills/twaldin/hone/slowest-tests)

Your own site

<a href="https://agentmods.dev/skills/twaldin/hone/slowest-tests"><img src="https://agentmods.dev/badge/skills/twaldin/hone/slowest-tests/github.svg" alt="Measured on agentmods" height="20"></a>

Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.

Your own site · 80×15

<a href="https://agentmods.dev/skills/twaldin/hone/slowest-tests"><img src="https://agentmods.dev/badge/skills/twaldin/hone/slowest-tests.svg" alt="Reviewed on agentmods" width="80" height="20"></a>

Per session 61 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.

When invoked 948 The whole file, excluding the scripts and references it only reads on demand.

Security scan A 0 findings. A grade says what 26 rules found in the file — not that it is safe.

Origin original No closer match found in the catalogue.

Token cost

What it costs to keep this loaded

Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.

Model	Per session	Once invoked
Fable 5.1	$0.00061	$0.00948
Opus 5	$0.00030	$0.00474
Sonnet 5	$0.00012	$0.00190
Haiku 4.5	$0.00006	$0.00095

Measured yesterday against content hash 64faac05cc20, method: parsed. Prices are Anthropic first-party input rates as of 2026-09-08, from the pricing page.

Security

Grade A, and why

slowest-tests scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured yesterday.

A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.

Nothing flagged

None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.

capsules/bun-module-loader/.candidate-5187e276/.claude/skills/slowest-tests/SKILL.md · 51 lines

How it starts

The opening of the file, as written. The whole thing — 51 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Slowest CI tests

Args: [N] [#channel] — both optional. N defaults to 500. If #channel is omitted, just print the table and stop.

1. Gather the data

Run the script. With no build number it auto-picks the most recent finished build from a merged PR.

bun run ci:slowest                 # top 500 from a recent merged-PR build → TSV on stdout
bun run ci:slowest 47324 100       # specific build, top 100
bun run ci:slowest --json > /tmp/slow.json

The script (scripts/ci-slowest-tests.ts) does the heavy lifting:

Lists test-bun jobs from bk build view <N>, skipping retried: true.
Fetches each job's raw_log_url directly with Authorization: Bearer $BUILDKITE_TOKEN. Do not use bk job log — it hangs indefinitely on some Windows/alpine jobs.
Caches logs to $TMPDIR/bun-ci-logs-<build>/ so re-runs are instant.
Parses _bk;t=<ms> ... --- [N/TOTAL] <file> group headers, normalising backslashes to / and stripping [attempt #N] retries.
Aggregates each file's duration as the max across all platforms (a file appears once per platform; shards within a platform are disjoint).
Drops package.json / non-JS entries — those are setup steps, not tests.

If the script can't find a build automatically (rare — it walks the last 10 merged PRs), pick one yourself: gh pr list --state merged --limit 10 --json number,headRefName, then bk build list --branch <headRefName> and pass the first build with a finished_at. Merged-PR builds usually report state: failed because of flaky tests — that's fine, the timing data is still valid.

2. Post to Slack (only if a channel was given)

slack_send_message does NOT support markdown tables — its markdown→blocks converter rejects | a | b | syntax with invalid_blocks. Don't use Canvases either; they render tables but the MCP proxy 502s above ~10 KB and the result is clunky.

Procedure:

Look up the channel ID with slack_search_channels (the user gives a name, you need the C… ID).
Write the full N-row markdown table to ~/code/tmp/top<N>-slow-tests.md, then upload it as a secret gist: gh gist create <file> --desc "Bun CI: top N slowest test files (build #<num>)". (The Slack MCP has no file-upload tool; secret gist is the agreed fallback. Do not ask the user to attach anything manually.)
Post the main message: header (build link + gist link, "Rest in thread.") followed by the top 20 bullets. Row format:

Read the full file on GitHub · 51 lines

Changes

What this file has done since we first saw it

Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.

yesterday First seen · 51 lines · 61 tokens per session scan A 64faac05cc20

Subscribe to this mod's changes

slowest-tests is a skill published in the GitHub repository twaldin/hone (47 stars, last pushed yesterday), licensed MIT. It adds 61 tokens to every session and 948 once invoked, about $0.0003 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-09-07.

DevOps and cloud Testing anthropic chatgpt claude cli coding-agent gepa

bug-fix

A structured bug-fixing workflow that takes a problem from reproduction and evidence-based cause finding through approval, implementation, testing, and cleanup.

rushengzhou/sid-code · 62 tokens

test-fix

Diagnose and minimally fix a failing test, then re-run.

terva-sh/terva · 17 tokens

evaluating-llms-harness

Evaluates LLMs across 60+ academic benchmarks (MMLU, HumanEval, GSM8K, TruthfulQA, HellaSwag). Use when benchmarking model quality, comparing models, reporting academic results, or tracking training progress. Industry standard used by EleutherAI, HuggingFace, and major labs. Supports HuggingFace, vLLM, APIs.

NousResearch/hermes-agent · 29 tokens

slowest-tests