Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Use BEFORE generating, refactoring, reviewing, or debugging code. Trigger phrases include "write a function/script/class for X", "review this code/diff/PR", "refactor this", "debug this error", "is this implementation correct", "what's wrong with this code", "improve this code", "translate from X to Y", or any prompt with a code block the user wants you to act on. Also fires when planning architectural changes, picking algorithms or data structures, or evaluating dependency upgrades. Calls the c
.claude/skills/jeremylongshore-code/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-06 | ✓→✗ | ▼ Worse | -86% | 0% |
| case-07 | ✓→✗ | ▼ Worse | -91% | 0% |
| case-08 | ✓→✗ | ▼ Worse | -87% | 0% |
| case-09 | ✓→✗ | ▼ Worse | -83% | 0% |
| case-10 | ✓→✗ | ▼ Worse | -80% | 0% |
When this skill triggers, call the code tool from the ejentum MCP server. Pass a 1-2 sentence framing of WHAT you are coding or reviewing as the query argument. Include the failure risk to avoid where possible.
Good query: review a Python refactor that converts raise UserNotFound to silent default return; tests still pass Bad query: look at this code
The tool returns a structured scaffold containing:
[CODE FAILURE]: engineering failure pattern to avoid[ENGINEERING PROCEDURE]: steps to follow[REASONING TOPOLOGY]: decision flow[CORRECT PATTERN]: shape correct code should take[VERIFICATION]: self-checkAmplify: and Suppress: signalsAbsorb internally. Do NOT echo bracket labels in the user-facing reply. Apply the scaffold's failure-pattern check against your draft before responding; if your code exhibits the named failure, rewrite.
If the API is unreachable, proceed with native engineering. The scaffold enhances; it is not a hard dependency.
Latency cost: ~1 second. Benefit: catches the kinds of behavioral changes and silent contract violations that look plausible but break under real conditions.
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→fail | 7,507 | 5,136 | -32% | 1 | 1 | 0% | 1,358 | 484 | -64% | 0 | 0 | — |
case-02 | fail→fail | 6,553 | 6,867 | +5% | 1 | 1 | 0% | 959 | 583 | -39% | 0 | 0 | — |
case-03 | fail→fail | 21,597 | 4,555 | -79% | 1 | 1 | 0% | 4,078 | 442 | -89% | 0 | 0 | — |
case-04 | fail→fail | 11,507 | 5,455 | -53% | 1 | 1 | 0% | 2,195 | 419 | -81% | 0 | 0 | — |
case-05 | fail→fail | 16,188 | 4,400 | -73% | 1 | 1 | 0% | 3,398 | 425 | -87% | 0 | 0 | — |
case-06 | pass→fail | 17,895 | 4,992 | -72% | 1 | 1 | 0% | 3,171 | 459 | -86% | 0 | 0 | — |
case-07 | pass→fail | 24,802 | 4,774 | -81% | 1 | 1 | 0% | 4,401 | 408 | -91% | 0 | 0 | — |
case-08 | pass→fail | 19,000 | 6,059 | -68% | 1 | 1 | 0% | 3,291 | 434 | -87% | 0 | 0 | — |
case-09 | pass→fail | 14,770 | 5,208 | -65% | 1 | 1 | 0% | 2,791 | 466 | -83% | 0 | 0 | — |
case-10 | pass→fail | 12,535 | 5,416 | -57% | 1 | 1 | 0% | 2,624 | 535 | -80% | 0 | 0 | — |
case-11 | pass→pass | 20,965 | 17,219 | -18% | 1 | 1 | 0% | 4,357 | 3,811 | -13% | 0 | 0 | — |
case-12 | fail→fail | 14,405 | 4,989 | -65% | 1 | 1 | 0% | 2,419 | 411 | -83% | 0 | 0 | — |
case-13 | fail→fail | 10,508 | 5,321 | -49% | 1 | 1 | 0% | 2,133 | 442 | -79% | 0 | 0 | — |
case-14 | fail→fail | 17,808 | 4,697 | -74% | 1 | 1 | 0% | 3,148 | 410 | -87% | 0 | 0 | — |
case-15 | fail→fail | 13,479 | 5,226 | -61% | 1 | 1 | 0% | 2,267 | 501 | -78% | 0 | 0 | — |
case-16 | fail→fail | 17,964 | 4,565 | -75% | 1 | 1 | 0% | 3,543 | 445 | -87% | 0 | 0 | — |
case-17 | fail→fail | 20,397 | 4,558 | -78% | 1 | 1 | 0% | 3,805 | 424 | -89% | 0 | 0 | — |
case-18 | fail→fail | 14,826 | 4,375 | -70% | 1 | 1 | 0% | 2,935 | 432 | -85% | 0 | 0 | — |
case-19 | fail→fail | 9,721 | 4,487 | -54% | 1 | 1 | 0% | 2,216 | 469 | -79% | 0 | 0 | — |
case-20 | pass→fail | 12,882 | 4,923 | -62% | 1 | 1 | 0% | 2,043 | 487 | -76% | 0 | 0 | — |
case-21 | pass→fail | 9,183 | 4,767 | -48% | 1 | 1 | 0% | 1,543 | 569 | -63% | 0 | 0 | — |
case-22 | pass→fail | 12,115 | 5,431 | -55% | 1 | 1 | 0% | 2,024 | 528 | -74% | 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 1 counted toward the lift figure. The other 21 produced results that are not comparable between the two arms, so they are excluded from the headline rather than averaged into it. A headline lift is not published for this run.
Other measured skills in the registry, with their headline benchmark lift.