Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Run full 5-layer analysis (AST, call graph, CFG, DFG, slice) on a specific function for deep debugging or understanding.
| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-05 | ✗→✓ | ▲ Improved | -22% | 0% |
| case-06 | ✗→✓ | ▲ Improved | -18% | 0% |
| case-07 | ✗→✓ | ▲ Improved | -45% | 0% |
| case-09 | ✗→✓ | ▲ Improved | -31% | 0% |
| case-13 | ✗→✓ | ▲ Improved | -32% | 0% |
Full 5-layer analysis of a specific function. Use when debugging or deeply understanding code.
/tldr-deep <function_name>| Layer | Purpose | Command | |-------|---------|---------| | L1: AST | Structure | tldr extract <file> | | L2: Call Graph | Navigation | tldr context <func> --depth 2 | | L3: CFG | Complexity | tldr cfg <file> <func> | | L4: DFG | Data flow | tldr dfg <file> <func> | | L5: Slice | Dependencies | tldr slice <file> <func> <line> |
Given a function name, run all layers:
bash# First find the file tldr search "def <function_name>" . # Then run each layer tldr extract <found_file> # L1: Full file structure tldr context <function_name> --project . --depth 2 # L2: Call graph tldr cfg <found_file> <function_name> # L3: Control flow tldr dfg <found_file> <function_name> # L4: Data flow tldr slice <found_file> <function_name> <target_line> # L5: Slice
## Deep Analysis: {function_name}
### L1: Structure (AST)
File: {file_path}
Signature: {signature}
Docstring: {docstring}
### L2: Call Graph
Calls: {list of functions this calls}
Called by: {list of functions that call this}
### L3: Control Flow (CFG)
Blocks: {N}
Cyclomatic Complexity: {M}
[Hot if M > 10]
Branches:
- if: line X
- for: line Y
- ...
### L4: Data Flow (DFG)
Variables defined:
- {var1} @ line X
- {var2} @ line Y
Variables used:
- {var1} @ lines [A, B, C]
- {var2} @ lines [D, E]
### L5: Program Slice (affecting line {target})
Lines in slice: {N}
Key dependencies:
- line X → line Y (data)
- line A → line B (control)
---
Total: ~{tokens} tokens (95% savings vs raw file)pythonfrom tldr.api import ( extract_file, get_relevant_context, get_cfg_context, get_dfg_context, get_slice ) # All layers for one function file_info = extract_file("src/processor.py") context = get_relevant_context("src/", "process_data", depth=2) cfg = get_cfg_context("src/processor.py", "process_data") dfg = get_dfg_context("src/processor.py", "process_data") slice_lines = get_slice("src/processor.py", "process_data", target_line=42)
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-02 | fail→fail | 29,537 | 11,691 | -60% | 1 | 1 | 0% | 6,189 | 3,124 | -50% | 0 | 0 | — |
case-01 | fail→fail | 5,085 | 3,998 | -21% | 1 | 1 | 0% | 682 | 1,549 | +127% | 0 | 0 | — |
case-03 | fail→fail | 22,327 | 6,273 | -72% | 1 | 1 | 0% | 3,507 | 1,042 | -70% | 0 | 0 | — |
case-04 | pass→pass | 8,737 | 3,147 | -64% | 1 | 1 | 0% | 1,457 | 1,432 | -2% | 0 | 0 | — |
case-05 | fail→pass | 11,351 | 4,176 | -63% | 1 | 1 | 0% | 1,905 | 1,493 | -22% | 0 | 0 | — |
case-06 | fail→pass | 10,271 | 2,880 | -72% | 1 | 1 | 0% | 1,664 | 1,360 | -18% | 0 | 0 | — |
case-07 | fail→pass | 14,112 | 2,367 | -83% | 1 | 1 | 0% | 2,324 | 1,287 | -45% | 0 | 0 | — |
case-08 | fail→fail | 11,677 | 3,278 | -72% | 1 | 1 | 0% | 1,954 | 1,520 | -22% | 0 | 0 | — |
case-09 | fail→pass | 10,198 | 3,888 | -62% | 1 | 1 | 0% | 2,082 | 1,439 | -31% | 0 | 0 | — |
case-10 | pass→pass | 2,945 | 2,431 | -17% | 1 | 1 | 0% | 496 | 1,151 | +132% | 0 | 0 | — |
case-11 | pass→pass | 17,518 | 2,735 | -84% | 1 | 1 | 0% | 2,679 | 1,344 | -50% | 0 | 0 | — |
case-12 | pass→pass | 4,283 | 2,235 | -48% | 1 | 1 | 0% | 635 | 1,206 | +90% | 0 | 0 | — |
case-13 | fail→pass | 10,930 | 2,649 | -76% | 1 | 1 | 0% | 1,921 | 1,303 | -32% | 0 | 0 | — |
case-14 | fail→fail | 20,124 | 1,972 | -90% | 1 | 1 | 0% | 3,581 | 1,076 | -70% | 0 | 0 | — |
case-15 | fail→fail | 12,970 | 2,456 | -81% | 1 | 1 | 0% | 1,989 | 1,272 | -36% | 0 | 0 | — |
case-16 | fail→pass | 6,841 | 4,561 | -33% | 1 | 1 | 0% | 1,171 | 1,341 | +15% | 0 | 0 | — |
case-17 | fail→pass | 8,036 | 2,905 | -64% | 1 | 1 | 0% | 1,222 | 1,350 | +10% | 0 | 0 | — |
case-18 | fail→pass | 7,652 | 1,629 | -79% | 1 | 1 | 0% | 1,375 | 1,091 | -21% | 0 | 0 | — |
case-19 | pass→pass | 9,042 | 7,379 | -18% | 1 | 1 | 0% | 1,925 | 2,046 | +6% | 0 | 0 | — |
case-20 | pass→pass | 18,714 | 19,713 | +5% | 1 | 1 | 0% | 3,860 | 4,803 | +24% | 0 | 0 | — |
case-21 | pass→pass | 9,149 | 8,072 | -12% | 1 | 1 | 0% | 1,733 | 2,284 | +32% | 0 | 0 | — |
case-22 | pass→pass | 13,552 | 11,492 | -15% | 1 | 1 | 0% | 2,011 | 2,577 | +28% | 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 +36 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.
| Model | Method | Date | Lift |
|---|---|---|---|
| gemini-3.6-flash | verified | 7/29/2026 | +59% |
Other measured skills in the registry, with their headline benchmark lift.