From autodesign
Analyze results from an autodesign exploration loop. This skill should be used when the user asks to "analyze design results", "show design progress", "summarize design exploration", "what designs worked", "design statistics", "which exploration axis performed best", or needs to review the results from design-results.tsv.
How this skill is triggered — by the user, by Claude, or both
Slash command
/autodesign:analyze-resultsThis skill is limited to the following tools:
The summary Claude sees in its skill listing — used to decide when to auto-load this skill
Analyze design-results.tsv from autodesign exploration loops to extract insights,
Analyze design-results.tsv from autodesign exploration loops to extract insights, identify patterns across exploration axes, and suggest next directions.
Read design-results.tsv (or path from design-config.yaml results_file field).
Parse tab-separated columns (6 columns):
| Index | Field | Description |
|---|---|---|
| 0 | commit | 7-char git hash |
| 1 | composite_score | Composite score (0-100) |
| 2 | status | keep/discard/constraint_fail/crash |
| 3 | constraint | Which constraint failed, or "-" |
| 4 | axis | Exploration axis used |
| 5 | description | Short description |
Calculate and report:
For each exploration axis, calculate:
| Axis | Tried | Kept | Rate | Best Delta | Avg Delta |
|---|
Identify which constraints fail with which axes.
Report which axes from config have been explored vs. unexplored. Flag axes with zero attempts as unexplored.
Based on the analysis, suggest:
Use the bundled analysis script for automated parsing:
python3 ${CLAUDE_PLUGIN_ROOT}/scripts/parse-results.py design-results.tsv higher
This outputs JSON with summary statistics and axis-level breakdown.
npx claudepluginhub hironow/dotfiles --plugin autodesignDocuments results of experiments or A/B tests with statistical analysis, learnings, and recommendations. Use after experiments conclude to communicate findings and inform decisions.
Runs rigorous statistical analysis for ML/AI experiments: validates artifacts, computes descriptive/inferential stats, generates scientific figures, and surfaces missing evidence.
Records Simmer iteration scores in trajectory table, tracks best candidate, handles regression rollback, and passes ASI to next round. Supports single-file and workspace modes.