Stats
Actions
Tags
From workflow-optimizer
Diff two metric snapshots (before vs after) and determine if a workflow improved, regressed, or stagnated. Standalone — does not invoke other skills.
How this skill is triggered — by the user, by Claude, or both
Slash command
/workflow-optimizer:compareThis skill is limited to the following tools:
The summary Claude sees in its skill listing — used to decide when to auto-load this skill
Two metric files (baseline or iteration results):
Two metric files (baseline or iteration results):
Read both files and extract: success rate, avg duration, failure distribution.
For each metric: delta = after - before
| Metric | Before | After | Delta | Better if |
|---|---|---|---|---|
| Success rate | Higher | |||
| Avg duration | Lower |
Improved = success rate increased, OR same rate with lower duration. Regressed = success rate decreased. Stagnant = no meaningful change in any metric.
==============================================================
COMPARISON: {workflow-name}
==============================================================
Metric Before After Delta
--------------------------------------------------------------
Success Rate 60.0% 80.0% +20.0%
Avg Duration 185.0s 120.0s -65.0s
--------------------------------------------------------------
IMPROVED
==============================================================
Failure Distribution:
TIMEOUT 3 → 1 (-2)
TIMING 2 → 0 (-2)
AUTH 1 → 1 ( 0)
==============================================================
npx claudepluginhub yihan2099/workflow-optimizer --plugin workflow-optimizerDiffs baseline and rerun output directories from repeated plan runs to verify build reproducibility, masking non-deterministic fields like timestamps and IDs; reports drifted files with per-file changed-field summaries.
Compares harness evaluation history: shows score trends, per-tier deltas, diminishing returns detection, grade projections, bilingual reports, and ASCII charts. Useful after 2+ evaluations.
Benchmarks shell metric command N times across 2-3 git refs or repo states, checks variance, computes deltas vs baseline, outputs reproducible TSV table and summary. For honest code variant comparisons.