Eda Method Implementer

EDA Method Implementer

When to use

Use this skill when moving from experiment design to actual method/code implementation.

Knowledge And Tool Interaction

1. Load the approved hypothesis and experiment-plan artifacts before touching code.
2. If the task needs shared KB/tool lookup before implementation planning, delegate that step to eda-context-accessor.
3. If eda-context-accessor reports maintain, update, overturn, or add, carry that recommendation into the implementation handoff so later validation/infra skills can act on it.
4. Write implementation decisions and validation commands into explicit artifacts so later validation and retro skills can reuse them.

Workflow

1. Load approved hypothesis and experiment plan artifacts.
2. Query tool catalog before adding new scripts:

python3 scripts/common/tool_catalog.py query <keyword1> <keyword2>

3. Define implementation contract:

• target files/modules,
• expected behavior delta,
• metrics to validate change.

4. Implement minimal viable patch first, then iterate.
5. Run required validation commands and link outputs.
6. Prepare versioned checkpoint handoff to git-version-control.

Outputs

1. implementation_plan.md
2. implementation_change_log.md
3. validation_command_log.md

Hard rules

1. Do not implement without linked hypothesis ID.
2. Keep each patch testable and reversible.
3. Record known side effects and unresolved risks.

Reference

Load when needed:

1. references/implementation-contract-checklist.md