Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Scan conductor/ directory for project direction, workflows, and task planning when present
.claude/skills/aiskillstore-use-conductor/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-17 | ✗→✓ | ▲ Improved | 166% | 0% |
| case-04 | ✗→✓ | ▲ Improved | 12% | 0% |
| case-07 | ✗→✓ | ▲ Improved | -41% | 0% |
| case-08 | ✗→✓ | ▲ Improved | -32% | 0% |
| case-09 | ✗→✓ | ▲ Improved | -25% | 0% |
Scan the conductor/ directory at project root for structured project management files that provide direction, workflows, and task planning context.
Use this skill when:
The conductor system uses this structure:
textconductor/ ├── product.md # Product vision and purpose ├── product-guidelines.md # Standards and conventions ├── tech-stack.md # Technology decisions ├── workflow.md # Task execution methodology ├── tracks.md # Index of active work tracks ├── setup_state.json # Setup progress state ├── code_styleguides/ # Language-specific style guides │ ├── general.md │ └── python.md └── tracks/ # Detailed track plans └── <track_name>/ ├── spec.md # Track specification ├── plan.md # Task checklist with progress └── metadata.json # Track metadata
| File | Purpose | |------|---------| | product.md | Product vision, target audience, core features | | product-guidelines.md | Naming conventions, quality standards, documentation rules | | tech-stack.md | Approved technologies and libraries | | workflow.md | TDD methodology, task workflow, commit guidelines | | tracks.md | High-level index of all work tracks | | tracks/<name>/plan.md | Detailed task checklist with [ ], [~], [x] status | | tracks/<name>/spec.md | Goals, scope, and success criteria for the track |
conductor/ directory exists at project rootproduct.md for visionproduct-guidelines.md for standardstracks.md for active work[~]), read the track's plan.md to find current tasksworkflow.mdIn plan.md files:
[ ] - Task not started[~] - Task in progress[x] - Task completed (may include commit SHA)When conductor files are present:
plan.md[~] when starting, [x] when doneproduct-guidelines.md and style guidesspec.md for what's in/out of scopeBefore starting work on a project:
textUser: "What should I work on next?" Claude: [Checks for conductor/ directory] [Reads tracks.md to find active track] [Reads tracks/<active>/plan.md to find next [ ] task] "According to the conductor plan, the next task is..."
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-12 | pass→pass | 6,869 | 2,856 | -58% | 1 | 1 | 0% | 1,243 | 1,251 | +1% | 0 | 0 | — |
case-17 | fail→pass | 2,476 | 1,594 | -36% | 1 | 1 | 0% | 387 | 1,030 | +166% | 0 | 0 | — |
case-01 | fail→fail | 1,810 | 4,280 | +136% | 1 | 1 | 0% | 277 | 951 | +243% | 0 | 0 | — |
case-02 | fail→fail | 2,588 | 6,042 | +133% | 1 | 1 | 0% | 298 | 1,037 | +248% | 0 | 0 | — |
case-18 | pass→pass | 6,993 | 3,024 | -57% | 1 | 1 | 0% | 1,197 | 1,186 | -1% | 0 | 0 | — |
case-03 | fail→fail | 3,588 | 5,507 | +53% | 1 | 1 | 0% | 527 | 992 | +88% | 0 | 0 | — |
case-04 | fail→pass | 6,185 | 2,340 | -62% | 1 | 1 | 0% | 1,013 | 1,135 | +12% | 0 | 0 | — |
case-05 | pass→pass | 3,112 | 1,764 | -43% | 1 | 1 | 0% | 433 | 1,029 | +138% | 0 | 0 | — |
case-06 | pass→pass | 2,810 | 1,939 | -31% | 1 | 1 | 0% | 370 | 1,023 | +176% | 0 | 0 | — |
case-07 | fail→pass | 10,146 | 1,982 | -80% | 1 | 1 | 0% | 1,748 | 1,033 | -41% | 0 | 0 | — |
case-08 | fail→pass | 9,625 | 1,203 | -88% | 1 | 1 | 0% | 1,420 | 969 | -32% | 0 | 0 | — |
case-09 | fail→pass | 8,042 | 1,615 | -80% | 1 | 1 | 0% | 1,233 | 928 | -25% | 0 | 0 | — |
case-10 | fail→pass | 6,137 | 1,648 | -73% | 1 | 1 | 0% | 862 | 943 | +9% | 0 | 0 | — |
case-11 | fail→pass | 12,245 | 2,088 | -83% | 1 | 1 | 0% | 1,982 | 1,128 | -43% | 0 | 0 | — |
case-13 | fail→pass | 13,461 | 4,687 | -65% | 1 | 1 | 0% | 2,210 | 1,608 | -27% | 0 | 0 | — |
case-14 | fail→pass | 8,088 | 2,432 | -70% | 1 | 1 | 0% | 1,268 | 1,153 | -9% | 0 | 0 | — |
case-15 | fail→pass | 12,846 | 2,550 | -80% | 1 | 1 | 0% | 2,038 | 1,114 | -45% | 0 | 0 | — |
case-16 | fail→pass | 7,913 | 1,856 | -77% | 1 | 1 | 0% | 1,219 | 1,072 | -12% | 0 | 0 | — |
case-19 | fail→pass | 11,721 | 1,639 | -86% | 1 | 1 | 0% | 1,849 | 1,022 | -45% | 0 | 0 | — |
case-20 | pass→pass | 6,022 | 4,594 | -24% | 1 | 1 | 0% | 1,052 | 1,538 | +46% | 0 | 0 | — |
case-21 | pass→pass | 7,129 | 10,609 | +49% | 1 | 1 | 0% | 1,264 | 1,994 | +58% | 0 | 0 | — |
case-22 | pass→pass | 6,413 | 7,949 | +24% | 1 | 1 | 0% | 1,241 | 2,213 | +78% | 0 | 0 | — |
case-23 | pass→pass | 5,831 | 2,397 | -59% | 1 | 1 | 0% | 1,026 | 1,096 | +7% | 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. 23 cases were attempted, and 22 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 +52 percentage points is the difference between those two pass rates over the 22 comparable cases. 1 case got worse with the skill loaded, and it is included in that figure.
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.