Run the gated test classes that plain pytest skips (slow Docker/sandbox tests, live model-provider API tests, flaky tests, trio variants).
$ npx skills add UKGovernmentBEIS/inspect_ai --skill slow-tests -a claude-codeProject install by default; add -g for ~/.claude/skills/.
$ gh skill install UKGovernmentBEIS/inspect_ai slow-tests --agent claude-codeProject scope by default; add --scope user for a personal install. Needs GitHub CLI 2.90.0 or later (public preview).
$ git clone --depth 1 https://github.com/UKGovernmentBEIS/inspect_ai.git skills-src && mkdir -p .claude/skills && cp -r skills-src/.agents/skills/slow-tests .claude/skills/slow-tests && rm -rf skills-srcUse ~/.claude/skills/ instead of .claude/skills for a personal install. The folder must contain SKILL.md.
Claude Code skills documentation · loads skills from .claude/skills/
Install the "slow-tests" agent skill from https://github.com/UKGovernmentBEIS/inspect_ai/tree/main/.agents/skills/slow-tests into .claude/skills/slow-tests/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "slow-tests", then confirm the skill loads.Claude Code copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
$skill-installer install https://github.com/UKGovernmentBEIS/inspect_ai/tree/main/.agents/skills/slow-testsType this inside Codex. $skill-installer <name> installs a curated skill from openai/skills. The installer writes to $CODEX_HOME/skills (default ~/.codex/skills). Restart Codex if the skill does not show up.
$ npx skills add UKGovernmentBEIS/inspect_ai --skill slow-tests -a codexProject install goes to .agents/skills/; add -g for ~/.codex/skills/.
$ gh skill install UKGovernmentBEIS/inspect_ai slow-tests --agent codexProject scope by default (.agents/skills/); add --scope user for a personal install.
$ git clone --depth 1 https://github.com/UKGovernmentBEIS/inspect_ai.git skills-src && mkdir -p .agents/skills && cp -r skills-src/.agents/skills/slow-tests .agents/skills/slow-tests && rm -rf skills-srcUse ~/.agents/skills/ instead of .agents/skills for a personal install.
Codex skills documentation · loads skills from .agents/skills/
Install the "slow-tests" agent skill from https://github.com/UKGovernmentBEIS/inspect_ai/tree/main/.agents/skills/slow-tests into .agents/skills/slow-tests/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "slow-tests", then confirm the skill loads.Codex copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
$ npx skills add UKGovernmentBEIS/inspect_ai --skill slow-tests -a cursorProject install goes to .agents/skills/; add -g for ~/.cursor/skills/.
$ gh skill install UKGovernmentBEIS/inspect_ai slow-tests --agent cursorProject scope by default (.agents/skills/); add --scope user for a personal install.
$ git clone --depth 1 https://github.com/UKGovernmentBEIS/inspect_ai.git skills-src && mkdir -p .cursor/skills && cp -r skills-src/.agents/skills/slow-tests .cursor/skills/slow-tests && rm -rf skills-srcUse ~/.cursor/skills/ instead of .cursor/skills for a personal install.
Cursor skills documentation · loads skills from .cursor/skills/, .agents/skills/, .claude/skills/, .codex/skills/
Install the "slow-tests" agent skill from https://github.com/UKGovernmentBEIS/inspect_ai/tree/main/.agents/skills/slow-tests into .cursor/skills/slow-tests/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "slow-tests", then confirm the skill loads.Cursor copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
$ gemini skills install https://github.com/UKGovernmentBEIS/inspect_ai.git --path .agents/skills/slow-tests--scope user (default) or --scope workspace; --path is the subfolder of the repo that holds the skill; --consent skips the security confirmation prompt.
$ npx skills add UKGovernmentBEIS/inspect_ai --skill slow-tests -a gemini-cliProject install goes to .agents/skills/; add -g for ~/.gemini/skills/.
$ gh skill install UKGovernmentBEIS/inspect_ai slow-tests --agent gemini-cliProject scope by default (.agents/skills/); add --scope user for a personal install.
$ git clone --depth 1 https://github.com/UKGovernmentBEIS/inspect_ai.git skills-src && mkdir -p .gemini/skills && cp -r skills-src/.agents/skills/slow-tests .gemini/skills/slow-tests && rm -rf skills-srcUse ~/.gemini/skills/ instead of .gemini/skills for a personal install, then run /skills reload.
Gemini CLI skills documentation · loads skills from .gemini/skills/, .agents/skills/
Install the "slow-tests" agent skill from https://github.com/UKGovernmentBEIS/inspect_ai/tree/main/.agents/skills/slow-tests into .gemini/skills/slow-tests/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "slow-tests", then confirm the skill loads.Gemini CLI copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
$ gh skill install UKGovernmentBEIS/inspect_ai slow-testsInstalls for Copilot at project scope by default; add --scope user for a personal install. Preview a skill first with gh skill preview. Needs GitHub CLI 2.90.0 or later (public preview).
$ npx skills add UKGovernmentBEIS/inspect_ai --skill slow-tests -a github-copilotProject install goes to .agents/skills/; add -g for ~/.copilot/skills/.
$ git clone --depth 1 https://github.com/UKGovernmentBEIS/inspect_ai.git skills-src && mkdir -p .github/skills && cp -r skills-src/.agents/skills/slow-tests .github/skills/slow-tests && rm -rf skills-srcUse ~/.copilot/skills/ instead of .github/skills for a personal install. Commit .github/skills so cloud agent and code review can use it.
GitHub Copilot skills documentation · loads skills from .github/skills/, .claude/skills/, .agents/skills/
Install the "slow-tests" agent skill from https://github.com/UKGovernmentBEIS/inspect_ai/tree/main/.agents/skills/slow-tests into .github/skills/slow-tests/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "slow-tests", then confirm the skill loads.GitHub Copilot copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
$ npx skills add UKGovernmentBEIS/inspect_ai --skill slow-tests -a opencodeOpenCode documents no install command of its own. Project install goes to .agents/skills/; add -g for ~/.config/opencode/skills/.
$ gh skill install UKGovernmentBEIS/inspect_ai slow-tests --agent opencodeProject scope by default (.agents/skills/); add --scope user for a personal install.
$ git clone --depth 1 https://github.com/UKGovernmentBEIS/inspect_ai.git skills-src && mkdir -p .opencode/skills && cp -r skills-src/.agents/skills/slow-tests .opencode/skills/slow-tests && rm -rf skills-srcUse ~/.config/opencode/skills/ instead of .opencode/skills for a personal install.
OpenCode skills documentation · loads skills from .opencode/skills/, .claude/skills/, .agents/skills/
Install the "slow-tests" agent skill from https://github.com/UKGovernmentBEIS/inspect_ai/tree/main/.agents/skills/slow-tests into .opencode/skills/slow-tests/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "slow-tests", then confirm the skill loads.OpenCode copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
slow-testsRun the gated test classes that plain pytest skips (slow Docker/sandbox tests, live model-provider API tests, flaky tests, trio variants).
Slow Tests is an agent skill from UKGovernmentBEIS/inspect_ai. Run the gated test classes that plain pytest skips (slow Docker/sandbox tests, live model-provider API tests, flaky tests, trio variants). Use when asked to run slow, live, api, or trio tests, when debugging one, or when a change touches model providers, sandbox or tool code, agents, or async plumbing and needs its gated coverage run.
Its SKILL.md is about 1.4k tokens, which your agent loads only when the skill is triggered. It is a single SKILL.md file with no bundled scripts.
It sits in Testing & QA, covering Unit testing, Containers and API testing. It works with Docker and pytest. The repository describes itself as: Inspect: A framework for large language model evaluations. The licence is MIT.
Read from SKILL.md and the folder at commit a1dbea8. It shows what the files ask for, not the result of running them.
Pre-approves nothing: there is no allowed-tools line, so your agent's usual permission prompts apply.
From allowed-tools in the SKILL.md frontmatter.
Shell commands in SKILL.md call:
pytestmakedockerFrom the folder's file list and the shell code blocks in SKILL.md.
No URLs in SKILL.md. Its commands use docker, which can reach the network depending on how they are called.
From URLs in SKILL.md, links to its own repository left out.
Names these keys or tokens, usually read from environment variables:
OPENAI_API_KEYANTHROPIC_API_KEYGOOGLE_API_KEYGROQ_API_KEYMOONSHOT_API_KEYTOGETHER_API_KEYFrom names ending in _API_KEY, _TOKEN, _SECRET, _KEY or _PASSWORD in SKILL.md.
Slow Tests loads about 1.4k tokens when it runs. Until then it costs about 87 tokens; SKILL.md has 675 words of instructions outside code blocks.
Estimates: characters ÷ 4, the usual rule of thumb; real counts depend on the model's tokenizer. Scripts and assets cost tokens only if the agent reads them.
The automated check found no risky patterns in SKILL.md.
Automated static check — not a guarantee. Review scripts before installing. It scans the text of SKILL.md for risky patterns (piping downloads into a shell, reading credential files, hidden Unicode, destructive commands); files beside SKILL.md are not scanned.
The full file from UKGovernmentBEIS/inspect_ai at commit a1dbea8, republished under its MIT licence (© UKGovernmentBEIS). 675 words, ~1,366 tokens.
.claude/skills/slow-tests/SKILL.md (or your agent's skills folder).Plain pytest (and make test) runs the fast suite only. Four classes of
test are gated behind flags defined in tests/conftest.py. A gated test
that is not enabled shows as skipped, and the run still exits green. Enable
the class, then read the skip summary at the end of the run (-ra is in
addopts) to confirm the tests you meant to run actually ran.
PR CI runs almost none of these. A scheduled job in meridianlabs-ai/actions
runs pytest --runslow --runapi against main every couple of hours, so a
gated test that a change breaks fails there, after merge.
| Class | Flag | How a test joins it | What it needs | Does PR CI run it? |
|---|---|---|---|---|
| slow | --runslow | @pytest.mark.slow | Usually Docker (@skip_if_no_docker), sometimes a provider key too | Only tests/tools/ when tool or sandbox code changed, plus the areas in the detect-slow table in .github/workflows/build.yml |
| api | --runapi | skip_if_no_<provider> decorators in tests/test_helpers/utils.py add the api marker | The provider's key or opt-in variable; the decorator names it | No. CI has no provider keys |
| flaky | --runflaky | @pytest.mark.flaky | Whatever the underlying test needs (Docker or a key) | No |
| trio | --runtrio | Every async test gets a [trio] variant automatically | A separate pytest invocation; it runs only the trio variants | No |
Flags combine, except --runtrio, which must be its own run.
Run the gated tests that cover the code you changed, scoped to the test directories that exercise it. The whole gated suite is large and spends provider credit; leave that to the scheduled job.
| You changed | Run |
|---|---|
src/inspect_ai/model/_providers/** or a provider conversion module in src/inspect_ai/model/ | pytest --runapi --runslow tests/model/providers/test_<provider>*.py with that provider's key set. Shared code (_providers/util/, openai_compatible.py, providers.py) is used by many providers; run tests/model tests/tools tests/agent with a key for each provider built on it |
litellm_proxy.py or _litellm_proxy_*.py in src/inspect_ai/model/_providers/ | pytest --runslow tests/model/providers/test_litellm_proxy*.py, then again with --runtrio. The Docker tests run the proxy image and fake upstreams offline. Add --runapi for the live round trips; each live deployment skips unless its key is set (see "LiteLLM proxy" below) |
src/inspect_ai/tool/** or src/inspect_sandbox_tools/** | pytest --runslow -m slow tests/tools/ -x with Docker running. Add --local-inspect-tools to build the sandbox tools from your working tree instead of using the published binary (see src/inspect_sandbox_tools/AGENTS.md) |
src/inspect_ai/util/_sandbox/** or src/inspect_ai/util/_checkpoint/** | pytest --runslow tests/util/ tests/checkpoint/ with Docker running |
src/inspect_ai/agent/** | pytest --runslow --runapi tests/agent/. Many of these are ACP and TUI tests that need only time; the rest need Docker or an OpenAI or Anthropic key |
Async plumbing (task groups, cancellation, anything under _util/_async.py or a bridge) | The affected tests once normally and once with --runtrio |
A test marked flaky, or the code it covers | That test with --runflaky |
Most providers gate on <PROVIDER>_API_KEY (OPENAI_API_KEY,
ANTHROPIC_API_KEY, GOOGLE_API_KEY, GROQ_API_KEY, and so on). Bedrock
and Vertex authenticate through AWS and gcloud credentials, so their gate is
an opt-in switch: ENABLE_BEDROCK_TESTS, ENABLE_VERTEX_TESTS. Azure-hosted
OpenAI and Mistral skip on their AZUREAI_* variables without the api
marker. When unsure, read the decorator on the test.
A provider file's tests often gate on several keys. With only
OPENAI_API_KEY set, test_openai*.py still reports skips for Together,
Bedrock, vLLM, and reasoning-summary access. Those are expected. A skip
whose reason names the key you set means that test did not run.
The litellm-proxy tests never pull images. They skip unless
ghcr.io/berriai/litellm:main-latest (override with LITELLM_PROXY_IMAGE)
is present locally, and the key and team alias tests also need
postgres:16 (LITELLM_PROXY_POSTGRES_IMAGE). Pull them once with
docker pull. The live tests pass the provider keys to the proxy container:
OPENAI_API_KEY, ANTHROPIC_API_KEY, GOOGLE_API_KEY, MOONSHOT_API_KEY,
TOGETHER_API_KEY, and ENABLE_BEDROCK_TESTS with AWS credentials.
@skip_if_no_docker checks that the docker binary exists, not that the
daemon is up. With Docker installed but stopped the tests fail instead of
skipping. Start Docker first.
Local runs set no --timeout (CI uses 900s); conftest wraps each async
test attempt in a 5-minute timeout, sync tests get none. If a test hangs,
rerun with --timeout=<seconds>; the conftest watchdog then dumps every
thread stack 300 seconds before the kill, or at the threshold in
INSPECT_TEST_HANG_DUMP_SECONDS if you set one.
© UKGovernmentBEIS, MIT. Rendered from Markdown: HTML in the file is shown as text, images as links, and headings moved down two levels. Raw file
Just SKILL.md in .agents/skills/slow-tests of UKGovernmentBEIS/inspect_ai.
Open the folder on GitHubat commit a1dbea8
Slow Tests next to the 5 skills that share the most tags, products or categories with it. Stars are the repository's; “used in” counts other GitHub owners with a copy.
| Skill | Stars | Used in | Tokens | Auto-check | Licence | Repo updated |
|---|---|---|---|---|---|---|
| Slow Tests this skillUKGovernmentBEIS/inspect_ai | 3k | — | ~1.4k | Automated safety check: Pass | MIT | |
| Flowfile Testing And ValidationEdwardvaneechoud/Flowfile | 385 | — | ~8.3k | Automated safety check: Notes | MIT | |
| Code PatternsAedelon/claude-code-blueprint | 120 | — | ~1.2k | Automated safety check: Pass | Custom licence | |
| Flowfile Debugging PlaybookEdwardvaneechoud/Flowfile | 389 | — | ~6.3k | Automated safety check: Pass | MIT | |
| Redamon Testingsamugit83/redamon | 3k | — | ~1.8k | Automated safety check: Pass | MIT | |
| Maintainer Testing ReleaseVectorSpaceLab/AREX-Skill | 331 | — | ~697 | Automated safety check: Pass | AGPL-3.0 |
Edwardvaneechoud/Flowfile
Exact per-package pytest/vitest/playwright commands, the registered pytest markers and which need Docker, the testutils Docker fixture matrix, the shared-test-DB isolation model and its failure…
Aedelon/claude-code-blueprint
Reference patterns for REST APIs, pytest/vitest testing, Docker multi-stage builds, GitHub Actions CI/CD, PostgreSQL, TypeScript generics, Python async, and React Server Components.
Edwardvaneechoud/Flowfile
Symptom-to-cause triage playbook for Flowfile (core/worker/kernel/frontend/AI) — covers "no such table" DB cascades (two distinct causes), import-time Alembic migration corruption, silent…
samugit83/redamon
How RedAmon tests actually run and how to author them: the per-file Docker gate, the unit/integration/live tiers, and the failure modes that make a green run a lie.
VectorSpaceLab/AREX-Skill
Use this quip-miner sub-skill for maintainer pytest selection, CI invariants, no-inline-sampling lint, multiprocessing/hang debugging, versioning, Docker/PyInstaller release checks, and safe…
microsoft/onnxruntime
Runs and debugs ONNX Runtime tests: Google Test executables for C++ and unittest or pytest for Python, with filters and build-directory guidance.
UKGovernmentBEIS/inspect_ai
Measure PR CI speed, queue and execution time, slow tests, suite growth, and runner waste.
UKGovernmentBEIS/inspect_ai
Land a PR that requires new inspect-sandbox-tools injectable binaries to be built and published.
UKGovernmentBEIS/inspect_ai
Land a PR that requires a coordinated ts-mono submodule change.
Categories
Run the gated test classes that plain pytest skips (slow Docker/sandbox tests, live model-provider API tests, flaky tests, trio variants). Slow Tests is an agent skill from UKGovernmentBEIS/inspect_ai. Run the gated test classes that plain pytest skips (slow Docker/sandbox tests, live model-provider API tests, flaky tests, trio variants).
Slow Tests fits situations like: asked to run slow; A change touches model providers; async plumbing and needs its gated coverage run.
Run `npx skills add UKGovernmentBEIS/inspect_ai --skill slow-tests -a claude-code`. Or copy the skill folder (.agents/skills/slow-tests in UKGovernmentBEIS/inspect_ai) into .claude/skills/slow-tests in your project. Claude Code loads it when a task matches its description.
Run `npx skills add UKGovernmentBEIS/inspect_ai --skill slow-tests -a codex`. Or copy the skill folder (.agents/skills/slow-tests in UKGovernmentBEIS/inspect_ai) into .agents/skills/slow-tests in your project. Codex loads it when a task matches its description.
Cursor, Gemini CLI, GitHub Copilot and OpenCode also load SKILL.md folders. With the skills CLI, run `npx skills add UKGovernmentBEIS/inspect_ai --skill slow-tests -a cursor` (or -a gemini-cli, github-copilot or opencode for the others). To copy it by hand, put the folder in .cursor/skills/slow-tests, .gemini/skills/slow-tests, .github/skills/slow-tests and .opencode/skills/slow-tests in your project.
Going by SKILL.md and its folder, Slow Tests needs the command-line tools its instructions call (pytest, make and docker) and credentials named OPENAI_API_KEY, ANTHROPIC_API_KEY, GOOGLE_API_KEY and GROQ_API_KEY. Our summary lists: Docker; A credential in OPENAI_API_KEY; A credential in ANTHROPIC_API_KEY.
SKILL.md contains no URLs. Its commands use docker, which can reach the network depending on how they are called. This is read from the text; nothing was executed.
Our automated static check of SKILL.md found no risky patterns, such as piping downloads into a shell, reading credential files or hidden Unicode. It is not a guarantee. Review the folder before installing.
Slow Tests is published under the MIT licence (the repository's licence). It allows redistribution, so the full SKILL.md is shown on this page.
About 1.4k tokens (SKILL.md is roughly 5.5k characters). Agents keep only the skill's name and description in context until a task matches; then they load SKILL.md in full.
Skills that share tags, products or a category with Slow Tests: Flowfile Testing And Validation (Edwardvaneechoud/Flowfile, 385 stars), Code Patterns (Aedelon/claude-code-blueprint, 120 stars), Flowfile Debugging Playbook (Edwardvaneechoud/Flowfile, 389 stars) and Redamon Testing (samugit83/redamon, 3k stars). The comparison table on this page puts their stars, adoption, token cost, safety result and licence side by side.
UKGovernmentBEIS (a GitHub organization) maintains it in UKGovernmentBEIS/inspect_ai, which has 2,974 GitHub stars. The repository holds 4 skills in this directory. The repository was last updated on October 10, 2026.
Source: UKGovernmentBEIS/inspect_ai on GitHub. Facts on this page come from the repository at the commit we read; the author's words are quoted as theirs.
