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.
git clone --depth 1 https://github.com/stbenjam/skillsawWrote 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.
[](https://agentmods.dev/rules/stbenjam/skillsaw/testing)<a href="https://agentmods.dev/rules/stbenjam/skillsaw/testing"><img src="https://agentmods.dev/badge/rules/stbenjam/skillsaw/testing/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.
<a href="https://agentmods.dev/rules/stbenjam/skillsaw/testing"><img src="https://agentmods.dev/badge/rules/stbenjam/skillsaw/testing.svg" alt="Reviewed on agentmods" width="80" height="20"></a>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 | $0.00011 | $0.00661 |
| Opus 5.5 | $0.00004 | $0.00264 |
| Sonnet 5.5 | $0.00002 | $0.00132 |
| Haiku 4.5 | $0.00001 | $0.00066 |
Grade A, and why
testing scanned grade A with 1 finding 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 1mo ago.
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.
Runs shell commandslowCapability
Expected in a hook, worth knowing in a rule or an instructions file.
`subprocess.CompletedProcess` stand-in. Spawning an interpreter per call cost How it starts
The opening of the file, as written. The whole thing — 65 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Testing
Integration Tests
Integration tests live in tests/test_integration.py and verify end-to-end
behavior by driving the CLI (skillsaw lint / fix) against realistic repo
fixtures.
Drive the CLI through run_cli
run_lint and the _run_fix helpers call run_cli from
tests/cli_runner.py, which invokes skillsaw.cli.main() in-process with
argv, stdout, stderr, env and cwd swapped, and returns a
subprocess.CompletedProcess stand-in. Spawning an interpreter per call cost
~160ms — more than half of it importing skillsaw — and dominated the suite.
Prefer run_cli for a new integration test.
Reach for a real subprocess.run only when the test depends on something a
shared interpreter cannot give it:
- Import isolation — the
--no-custom-rulestests prove a rule file on disk did or did not get imported. In-process they would execute an arbitrary rule inside the test worker, andsys.modulescould mask a repeat import and make the negative case pass vacuously. - A real encoding on stdout — the lone-surrogate regressions assert that
rendering a report raises no
UnicodeEncodeError. AStringIOnever encodes, so in-process they would pass vacuously. - TTY and colour-cascade behavior, which depends on the real stream.
- Entry-point discovery —
tests/test_integration_plugins.pyinstalls plugins and needs a fresh interpreter to pick them up, so it stays on subprocess throughout.
Anything run_cli runs shares one interpreter, so a helper that leaves
process-global state behind breaks later tests. run_cli already resets the
two that outlive a call — root logging handlers and the skillsaw.utils file
cache; add to _reset_process_globals if you introduce another.
Use test fixtures
Prefer static fixtures in tests/fixtures/ over building repos
programmatically in test code. Fixtures should contain realistic file content
that mirrors what real users would write.
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.
- 1mo ago First seen · 65 lines · 11 tokens per session scan A 2b2164b556f5
testing is a cursor rule published in the GitHub repository stbenjam/skillsaw (67 stars, last pushed today), licensed Apache-2.0. It adds 11 tokens to every session and 661 once invoked, about $0.0000 per session on Opus 5.5. A static security scan graded it A with 1 finding (runs shell commands). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.
Other cursor rules, from other repositories
gemini_qa_loop
Continuous dogfood loop using Gemini 3 Flash vision to score and fix UI/UX issues automatically.
hatch3r-testing
Coverage thresholds, mocking strategy, property-based testing, mutation-score targets, flaky test quarantine, and snapshot test discipline.
ponytail
Ponytail, lazy senior dev mode. Always pick the simplest solution that works.
angular-20
This rule provides comprehensive best practices and coding standards for Angular development, focusing on modern TypeScript, standalone components, signals, and performance optimizations.
dev-standard
Apache Superset development standards and guidelines for Cursor IDE.
prefer-direct-imports-over-module-mocks
Prefer extracting a testable core over vi.mock / vi.resetModules when unit tests need to reach production logic entangled with config, env, or singletons.