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/coddy-project/coddy-agentWrote 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/coddy-project/coddy-agent/testing)<a href="https://agentmods.dev/rules/coddy-project/coddy-agent/testing"><img src="https://agentmods.dev/badge/rules/coddy-project/coddy-agent/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/coddy-project/coddy-agent/testing"><img src="https://agentmods.dev/badge/rules/coddy-project/coddy-agent/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 | $0.02616 | $0.02616 |
| Opus 5.5 | $0.01046 | $0.01046 |
| Sonnet 5.5 | $0.00523 | $0.00523 |
| Haiku 4.5 | $0.00262 | $0.00262 |
Grade A, and why
testing 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 today.
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.
How it starts
The opening of the file, as written. The whole thing — 56 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Testing (Go)
Commands
- Express run (what agents run before finishing):
make test- Runs
make test-agent-rules,make ui-build,make ui-test(the SPA's vitest suite), then onego test -tags=http,ui,scheduler,memory,cli,gateway,swarm,metrics ./...- the whole tree with every optional module compiled in, which is what the shipped binary contains. Minutes, not tens of minutes.
- Runs
- Tag matrix (CI's job, not a local step):
make test-matrixwalks every combination inTEST_TAG_SETS(Makefile) - the untagged build with its stubs, single tags, pairs, the shipped set..github/workflows/tests-on-pr.yamlreads that list throughmake print-test-tag-setsand runs one job per combination withfail-fast: false, so a broken stub is named by the job that failed. Locally, a build-tag boundary change (a_stub.go, anAvailableconst, a//go:buildline) is checked with that one combination:go test -tags=<set> ./.... The untagged build is also compiled bymake lint, so a signature a stub no longer satisfies is caught before the push. - Race detector:
make test-raceruns the whole tree undergo test -racewith every tag butui; CI runs it on every pull request (job Race detector). Run it before the push when a change touches goroutines, locks or channels, or a test harness that drives a live loop (the console's UI loop, a runningApp, a godog step reading state a worker writes).GOFLAGS=-count=3 make test-racerepeats every test, which a test that leans on process-wide state (a counter,bgtask.Default()) has to survive. Every package is clean under it: a race it reports is fixed where it lives (a stub's field read under the stub's own lock, a channel handed to a handler goroutine as a local), never skipped. - Prompt-cache group:
make test-cacheruns thefeatures/prompt_cache_*.featurespecs and every test namedTestPromptCache*- what each request sends the provider: one system message for the whole session, a history that only grows at its end, every rule at most once and none that no path brought in. It is part ofmake test; run it on its own when a change touches the system prompt, the rules, the history or the send boundary, and give a new test of that kind theTestPromptCacheprefix so it joins the group. - Performance group:
make test-perfruns the benchmarks (Benchmark*in theperf_test.gofiles), whichmake testnever runs: timings of the work a turn repeats and the context a session over this repository's own rules costs (system-tokens,request-tokens,duplicate-rulesmetrics).BENCH=<regexp>narrows them,BENCHTIME=5xsets the runs. A benchmark goes into the package'sperf_test.go, not into its functional tests. - Targeted run while iterating:
go test ./path/to/pkg -run TestName -count=1 - Touching a file behind
//go:build windows, or any signature it shares with_other.go/ the rest of the tree: also runmake check-windowsandmake lint-windows.make testandmake lintrun on the host, so the Windows half ofinternal/platform(and anything else with a per-OS file) is never compiled by them.check-windowscross-builds andgo vets every non-uitag combination, test files included. CI runs both (.github/workflows/tests-on-pr.yaml), plusgo teston a realwindows-latestrunner forinternal/platform,internal/bgtask,internal/tools/shell,internal/update,internal/hooks,internal/rules(drive-letter glob anchoring) and the console TUI (external/cli,-tags=cli) — the packages whose behavior differs by OS — and forcmd/coddy,internal/config,internal/dryrunandinternal/swarm, whose tests write files and compare paths. The platform packages and the console suite run onmacos-latestas well (jobtest-macos), which also builds the console and drives the binary through a real pty withexamples/cli/cli_e2e_startup.py; theclijob of the Linux test matrix runs that script too, and the Telegram stand (tgfake) job runsexamples/gateway/tg_e2e_offline.sh: the built binary against the release binary of tgfake (the versiongo.modpins, installed by tgfake's own action; guidedocs/contributing/tgfake.md), no Telegram and no model. The Go suite never opens a pty, so this is the only check of the console's terminal path. - Touching a file behind
//go:build android, orinternal/platform/android.gothat it calls: also runmake check-android. The Android build (GOOS=androidwith cgo, linked by the NDK against Bionic for arm64 and x86_64, published for Termux) compilesandroid_init.goand the cgo guardandroid_nocgo.go, which no host gate sees;check-androidrunsgo vetover it with the NDK compiler (scripts/android-cc.sh), test files included, and CI runs it in the Android cross-build job, which also builds both release binaries, checks that each is a position-independent executable naming/system/bin/linker64with no library Android does not ship, and uploads them as artifacts. The decisions themselves are plain functions tested on every Unix host (features/android_termux.feature,internal/platform/android_test.go). A new place that builds anexec.Cmdcallsplatform.AdaptCommandbeforeStart;TestEverySpawnSiteAdaptsTheCommandfails otherwise. - HTTP server core is
//go:build http. Embedded UI tests compile withgo test -tags=http,ui ./external/httpserver. SPA-freehttpbuild uses//go:build http && !uihandlers underexternal/httpserver. Session memory REST uses//go:build http && memory(memory_http.go); withoutmemory,memory_http_stub.goregisters no routes.
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.
- today Changed · +2 tokens per session 49a8795929aa
- 8d ago Changed · +31 tokens per session 53f9c1a50fde
- 10d ago Changed · +4 lines · +432 tokens per session dea3aa2497a6
- 11d ago Changed · +7 lines · +157 tokens per session 3b0b9523755c
- 18d ago Changed · +1 lines · +233 tokens per session 4b2c83f62df4
- 19d ago Changed · +2 lines · +263 tokens per session 8a3042fe8fda
- 25d ago Changed · +1 lines · +209 tokens per session fcef8ff0b710
- 26d ago Changed · +1 lines · +181 tokens per session 333ce13a5e48
- 27d ago Changed · +15 tokens per session b0d921b548be
- 28d ago First seen · 40 lines · 1,093 tokens per session scan A bf374a2c7af5
testing is a cursor rule published in the GitHub repository coddy-project/coddy-agent (175 stars, last pushed today), licensed MIT. It adds 2,616 tokens to every session, about $0.0105 per session on Opus 5.5. 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 2026-08-30.
Other cursor rules, from other repositories
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-assertions-over-defensive-checks
Prefer assertions over defensive checks when data is guaranteed to be valid.
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.
shared-plane-packages
Pattern for creating shared plane packages that serve both migration and runtime planes.