Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Guidelines for systematically exploring and understanding directory structures.
.claude/skills/esurovtsev-filesystem-navigation/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-08 | ✗→✓ | ▲ Improved | 146% | 0% |
| case-02 | ✓→✓ | ＝ Same ✓ | 23% | 0% |
| case-03 | ✓→✓ | ＝ Same ✓ | 90% | 0% |
| case-04 | ✓→✓ | ＝ Same ✓ | 31% | 0% |
| case-05 | ✓→✓ | ＝ Same ✓ | 113% | 0% |
When you are asked to explore, understand, or map a project's file structure. This includes requests like "what is this project?", "show me the structure", or "help me find where X is".
Always begin by listing the top-level directory. The root reveals the project type faster than anything else:
README.md or README.rst → start here, it's the author's own summaryrequirements.txt, pyproject.toml, package.json → tells you the language and dependenciesDockerfile, docker-compose.yml → the project is containerizedMakefile, justfile → there are predefined commands to run.env.example → environment variables are needed; never read .env itselfList all top-level directories before diving into any single one. Build a mental map:
src/ or app/ → application code lives heretests/ or test/ → test suiteconfig/ or conf/ → configurationdocs/ → documentationscripts/ or bin/ → utility scriptsmigrations/ or alembic/ → database migrationsDon't read every file. Choose what to read based on what you're trying to answer:
src/ recursivelyutils.py could contain anything.env files — they may contain secretsnode_modules/, __pycache__/, .git/, or other generated directories__main__.py or if __name__ == "__main__" block indicates an entry pointapp.py, main.py, or server.py is usually the entry point__init__.py files in Python indicate a package; they may re-export key symbols.gitignore, .flake8, .pre-commit-config.yaml) reveal tooling choices| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→fail | 3,684 | 4,360 | +18% | 1 | 1 | 0% | 219 | 861 | +293% | 0 | 0 | — |
case-02 | pass→pass | 5,715 | 4,206 | -26% | 1 | 1 | 0% | 1,179 | 1,449 | +23% | 0 | 0 | — |
case-03 | pass→pass | 2,832 | 2,083 | -26% | 1 | 1 | 0% | 522 | 993 | +90% | 0 | 0 | — |
case-04 | pass→pass | 3,851 | 2,516 | -35% | 1 | 1 | 0% | 830 | 1,087 | +31% | 0 | 0 | — |
case-05 | pass→pass | 2,595 | 2,333 | -10% | 1 | 1 | 0% | 464 | 990 | +113% | 0 | 0 | — |
case-06 | pass→pass | 7,818 | 3,766 | -52% | 1 | 1 | 0% | 1,513 | 1,321 | -13% | 0 | 0 | — |
case-07 | pass→pass | 6,179 | 3,859 | -38% | 1 | 1 | 0% | 1,342 | 1,322 | -1% | 0 | 0 | — |
case-08 | fail→pass | 2,543 | 2,213 | -13% | 1 | 1 | 0% | 402 | 987 | +146% | 0 | 0 | — |
case-09 | pass→pass | 2,147 | 2,353 | +10% | 1 | 1 | 0% | 395 | 1,024 | +159% | 0 | 0 | — |
case-10 | pass→pass | 7,160 | 4,369 | -39% | 1 | 1 | 0% | 1,409 | 1,429 | +1% | 0 | 0 | — |
case-11 | pass→pass | 6,797 | 4,935 | -27% | 1 | 1 | 0% | 1,336 | 1,539 | +15% | 0 | 0 | — |
case-12 | pass→pass | 3,420 | 2,124 | -38% | 1 | 1 | 0% | 678 | 987 | +46% | 0 | 0 | — |
case-13 | pass→pass | 5,477 | 4,243 | -23% | 1 | 1 | 0% | 1,189 | 1,368 | +15% | 0 | 0 | — |
case-14 | pass→pass | 8,897 | 5,503 | -38% | 1 | 1 | 0% | 1,413 | 1,714 | +21% | 0 | 0 | — |
case-15 | pass→pass | 4,602 | 2,829 | -39% | 1 | 1 | 0% | 804 | 1,231 | +53% | 0 | 0 | — |
case-16 | pass→pass | 2,414 | 1,804 | -25% | 1 | 1 | 0% | 440 | 946 | +115% | 0 | 0 | — |
case-17 | pass→pass | 2,043 | 2,797 | +37% | 1 | 1 | 0% | 411 | 1,033 | +151% | 0 | 0 | — |
case-18 | pass→pass | 1,532 | 1,445 | -6% | 1 | 1 | 0% | 257 | 826 | +221% | 0 | 0 | — |
case-19 | pass→pass | 1,669 | 3,634 | +118% | 1 | 1 | 0% | 292 | 1,173 | +302% | 0 | 0 | — |
case-20 | pass→pass | 2,404 | 2,133 | -11% | 1 | 1 | 0% | 373 | 974 | +161% | 0 | 0 | — |
case-21 | pass→pass | 2,046 | 1,773 | -13% | 1 | 1 | 0% | 359 | 892 | +148% | 0 | 0 | — |
case-22 | pass→pass | 4,206 | 2,490 | -41% | 1 | 1 | 0% | 627 | 1,066 | +70% | 0 | 0 | — |
DecimalAI ran this skill against gemini-3.6-flash twice over the same eval suite — once with the skill loaded and once without — and compared the two runs case by case. 22 cases were attempted, and 21 counted toward the lift figure. The other 1 produced results that are not comparable between the two arms, so they are excluded from the headline rather than averaged into it. The headline lift of +5 percentage points is the difference between those two pass rates over the 21 comparable cases.
Without the skill loaded, the model failed this case. With it loaded, the same prompt on the same model passed. This is one improved case from the latest verified run; every case, including any that regressed, is in the table above.
Other measured skills in the registry, with their headline benchmark lift.