Stats
Actions
Tags
From workflow-optimizer
Persist and load metric snapshots for tracking workflow optimization progress over time. Standalone — does not invoke other skills.
How this skill is triggered — by the user, by Claude, or both
Slash command
/workflow-optimizer:baselineThis skill is limited to the following tools:
The summary Claude sees in its skill listing — used to decide when to auto-load this skill
Save current measurement results as a baseline snapshot.
Save current measurement results as a baseline snapshot.
/baseline my-workflow save
Reads the latest results from .workflow-optimizer/{workflow-id}/baseline.md and copies to .workflow-optimizer/{workflow-id}/snapshots/{date}.md.
Load the most recent snapshot.
/baseline my-workflow load
Reads the latest file from .workflow-optimizer/{workflow-id}/snapshots/.
List all saved snapshots for a workflow.
/baseline my-workflow list
Output:
Snapshots for my-workflow:
2026-02-25.md (60.0% success, 5 runs)
2026-02-26.md (70.0% success, 5 runs)
2026-02-28.md (80.0% success, 5 runs)
.workflow-optimizer/
{workflow-id}/
baseline.md # Current metrics (written by measure/optimize)
snapshots/
{date}.md # Historical snapshots
iterations/
iter-1.md # Optimization iteration details
iter-2.md
failure-catalog.md # Accumulated failure patterns + fixes
# Snapshot: {workflow-id} — {date}
| Metric | Value |
|--------|-------|
| Runs | 5 |
| Success Rate | 80.0% |
| Avg Duration | 120.0s |
## Failure Distribution
| Category | Count |
|----------|-------|
| TIMEOUT | 1 |
## Run Details
| # | Success | Duration | Error |
|---|---------|----------|-------|
| 1 | PASS | 95s | |
| 2 | PASS | 130s | |
| 3 | FAIL | 300s | TIMEOUT |
| 4 | PASS | 110s | |
| 5 | PASS | 100s | |
npx claudepluginhub yihan2099/workflow-optimizer --plugin workflow-optimizerManages performance baseline JSON files by version, storing metrics and environment metadata. Ensures one file per version for consolidating results and comparing versions.
Metric-driven optimization loop in isolated worktrees: proposes changes, measures with a scalar metric command, keeps improvements, discards failures. Supports convergence detection and diminishing returns.
Runs autonomous optimization loops with 3 parallel agents per round, using shell metric commands, git worktrees, guards, and baselines to iteratively improve code performance.