Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Guides implementation of code tasks using test-driven development in an Explore, Plan, Code, Commit workflow. Acts as a Technical Implementation Partner and TDD Coach — following existing patterns, avoiding over-engineering, and producing idiomatic, modern code.
.claude/skills/mikeyobrien-code-assist/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-05 | ✗→✓ | ▲ Improved | 80% | 0% |
| case-08 | ✗→✓ | ▲ Improved | 120% | 0% |
| case-16 | ✗→✓ | ▲ Improved | 15% | 0% |
| case-12 | ✗→✓ | ▲ Improved | 42% | 0% |
| case-14 | ✗→✓ | ▲ Improved | 79% | 0% |
Guides implementation of code tasks using test-driven development in an Explore, Plan, Code, Commit workflow. Balances automation with user collaboration while adhering to existing package patterns and prioritizing readability and extensibility.
.code-task.md, or URLConstraints:
Apply these patterns throughout all steps:
Interactive Mode: Present actions for confirmation. Explain pros/cons when multiple approaches exist. Ask clarifying questions. Pause at key decision points. Provide educational context.
Auto Mode: Execute autonomously. Document all decisions and reasoning in progress.md. Select the most appropriate approach and document why. Provide comprehensive summaries at completion.
Separation of Concerns: Documentation goes in {documentation_dir}. Code (tests and implementation) goes in repo_root. Never mix them. Documentation should guide implementation with high-level concepts — not provide it. When including code snippets in documentation, keep them brief and clearly label them as examples or references.
CODEASSIST.md Integration: If CODEASSIST.md exists in repo_root, read it and apply its constraints throughout.
Initialize the project environment and create necessary directory structures.
Constraints:
{documentation_dir}/implementation/{task_name}/ (with logs subdirectory) and verify it exists before proceeding.code-task.md with YAML frontmatter, update status: in_progress and started: <date> (if not already set)> 💬 See Mode Behavior for mode-specific interaction guidance
Analyze the task description and existing documentation to identify core functionality, edge cases, and constraints.
Constraints:
> 💬 See Mode Behavior for mode-specific interaction guidance
Search for similar implementations and identify interfaces, libraries, and components the implementation will interact with.
Constraints:
> 💬 See Mode Behavior for mode-specific interaction guidance
Create test scenarios covering normal operation, edge cases, and error conditions.
Constraints:
> 💬 See Mode Behavior for mode-specific interaction guidance
Outline the high-level structure and create an implementation plan.
Constraints:
> 💬 See Mode Behavior for mode-specific interaction guidance
Phase-wide constraints:
{documentation_dir}/implementation/{task_name}/logs/ and search for success/failure indicatorsWrite test cases following strict TDD principles.
Constraints:
> 💬 See Mode Behavior for mode-specific interaction guidance
Write implementation code to pass the tests, focusing on simplicity and correctness first.
Constraints:
> 💬 See Mode Behavior for mode-specific interaction guidance
Review the implementation for simplification, improvement, and convention alignment.
Constraints:
> 💬 See Mode Behavior for mode-specific interaction guidance
Verify the implementation is complete and correct.
Constraints:
> 💬 See Mode Behavior for mode-specific interaction guidance
Draft a conventional commit message and perform the git commit.
Constraints:
.code-task.md, update frontmatter to status: completed and completed: <date> before committing> 💬 See Mode Behavior for mode-specific interaction guidance
{documentation_dir}/implementation/{task_name}/ (context.md, plan.md, progress.md, logs/)Input:
task_description: "Create a utility function that validates email addresses"
mode: "interactive"Expected Process:
Directory/Permission Issues: Create directories if possible, inform user of permission issues, suggest alternatives.
Build Failures: Check CODEASSIST.md for guidance, verify correct directory, try clean builds, check dependencies.
Multi-Package Coordination: Verify dependency order, build in order, create separate commits per package.
Task File Frontmatter Issues: Skip frontmatter updates if missing/malformed. Don't fail the task due to frontmatter issues.
Implementation Challenges: Document in progress.md, propose alternatives. In interactive mode ask for guidance; in auto mode select the most promising approach and document the decision.
• {documentation_dir}/implementation/{task_name}/context.md — Workspace structure, requirements, patterns, dependencies, implementation paths • {documentation_dir}/implementation/{task_name}/plan.md — Test scenarios, implementation strategy • {documentation_dir}/implementation/{task_name}/progress.md — Execution tracking, TDD cycles, commit status, challenges • {documentation_dir}/implementation/{task_name}/logs/ — Build outputs
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→fail | 5,019 | 6,016 | +20% | 1 | 1 | 0% | 270 | 2,456 | +810% | 0 | 0 | — |
case-22 | pass→pass | 10,499 | 6,590 | -37% | 1 | 1 | 0% | 1,996 | 3,269 | +64% | 0 | 0 | — |
case-07 | pass→pass | 5,116 | 2,566 | -50% | 1 | 1 | 0% | 1,006 | 2,585 | +157% | 0 | 0 | — |
case-02 | fail→fail | 2,645 | 7,572 | +186% | 1 | 1 | 0% | 345 | 2,517 | +630% | 0 | 0 | — |
case-03 | fail→fail | 4,456 | 5,305 | +19% | 1 | 1 | 0% | 253 | 2,501 | +889% | 0 | 0 | — |
case-04 | pass→pass | 3,662 | 3,802 | +4% | 1 | 1 | 0% | 771 | 2,809 | +264% | 0 | 0 | — |
case-05 | fail→pass | 7,186 | 1,955 | -73% | 1 | 1 | 0% | 1,337 | 2,405 | +80% | 0 | 0 | — |
case-06 | pass→pass | 9,837 | 4,345 | -56% | 1 | 1 | 0% | 1,900 | 2,842 | +50% | 0 | 0 | — |
case-08 | fail→pass | 6,250 | 2,872 | -54% | 1 | 1 | 0% | 1,115 | 2,448 | +120% | 0 | 0 | — |
case-09 | pass→pass | 6,271 | 2,325 | -63% | 1 | 1 | 0% | 1,126 | 2,426 | +115% | 0 | 0 | — |
case-10 | pass→pass | 10,510 | 4,509 | -57% | 1 | 1 | 0% | 1,922 | 2,847 | +48% | 0 | 0 | — |
case-11 | pass→pass | 5,881 | 3,174 | -46% | 1 | 1 | 0% | 1,004 | 2,649 | +164% | 0 | 0 | — |
case-16 | fail→pass | 11,149 | 3,194 | -71% | 1 | 1 | 0% | 2,338 | 2,688 | +15% | 0 | 0 | — |
case-12 | fail→pass | 13,131 | 5,892 | -55% | 1 | 1 | 0% | 2,210 | 3,149 | +42% | 0 | 0 | — |
case-13 | pass→pass | 6,617 | 2,610 | -61% | 1 | 1 | 0% | 1,201 | 2,514 | +109% | 0 | 0 | — |
case-14 | fail→pass | 7,390 | 1,042 | -86% | 1 | 1 | 0% | 1,207 | 2,160 | +79% | 0 | 0 | — |
case-15 | pass→pass | 7,684 | 1,247 | -84% | 1 | 1 | 0% | 1,556 | 2,176 | +40% | 0 | 0 | — |
case-17 | fail→pass | 6,608 | 2,897 | -56% | 1 | 1 | 0% | 1,190 | 2,612 | +119% | 0 | 0 | — |
case-18 | pass→pass | 3,610 | 1,622 | -55% | 1 | 1 | 0% | 610 | 2,194 | +260% | 0 | 0 | — |
case-19 | fail→pass | 12,232 | 3,658 | -70% | 1 | 1 | 0% | 2,283 | 2,744 | +20% | 0 | 0 | — |
case-20 | fail→fail | 10,934 | 13,966 | +28% | 1 | 1 | 0% | 624 | 3,519 | +464% | 0 | 0 | — |
case-21 | pass→fail | 5,928 | 7,796 | +32% | 1 | 1 | 0% | 1,193 | 3,396 | +185% | 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 19 counted toward the lift figure. The other 3 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 +27 percentage points is the difference between those two pass rates over the 19 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.