Performance Regression Verification
Use when someone claims a performance improvement or worries about a regression and needs proof. Covers before-versus-after comparison, stable verification scenarios, user-visible metrics, and guarding against “faster on my machine” conclusions.
Install
$ agentstack add skill-andrewsrigom-agent-skills-performance-regression-verification ✓ scanned · ✓ verified, works with Claude Code, Cursor, and more.
Security review
✓ PassedNo issues found. Passed automated security review. · v0.1.0 How review works →
- ✓ Prompt-injection patterns
- ✓ Secret / credential exfiltration
- ✓ Dangerous shell & filesystem operations
- ✓ Untrusted network calls
- ✓ Known-malicious package signatures
What it can access
- ✓ Network access No
- ✓ Filesystem access No
- ✓ Shell / process execution No
- ✓ Environment & secrets No
- ✓ Dynamic code execution No
From automated source analysis of v0.1.0. “Used” means the capability is present in the source — more access means more to trust, not that it’s unsafe.
Verified badge
Passed review? Show it. Paste this badge into your README, it links to the public security report.
Links to your public security report.[](https://agentstack.voostack.com/security/report/skill-andrewsrigom-agent-skills-performance-regression-verification) Reliability & compatibility
Declared compatibility
Compatibility is declared by the source manifest. End-to-end runtime verification is coming, see below.
We're building live execution health for every listing: tool-call success rate, median latency, uptime, and last-checked timestamps, measured, not self-reported. It isn't live yet, so we don't show numbers we can't stand behind.
How agent discovery & health will work →About
Performance Regression Verification
Use this skill when the job is proving performance got better, stayed stable, or regressed in a meaningful way.
Scope
- before and after comparison
- guarding against accidental slowdowns
- verifying optimizations before shipping
- choosing the right metric and acceptance bar
- separating real improvement from benchmark noise
Default path
- Name the scenario being protected.
- Choose the metric that matters for that scenario.
- Compare the same path before and after.
- Check both user-visible improvement and correctness.
- Report confidence and residual risk instead of pretending performance is binary.
When to deviate
- Use percentile data when tail latency matters more than averages.
- Use smoke thresholds rather than exact equality when CI noise is unavoidable.
- Prefer field telemetry over lab checks when synthetic runs miss the real pain.
Guardrails
- Compare the same scenario, environment, and data shape when possible.
- Do not call a change “faster” without saying what metric improved.
- Do not treat tiny wins as meaningful if the user-visible bottleneck remains.
- Keep correctness, stability, and resource usage in the verification story.
Avoid
- “feels faster” as the only evidence
- comparing different inputs or environments
- only reporting averages when tail latency is the real issue
- dropping regression checks once the optimization ships
Verification checklist
- the protected scenario is explicit
- before and after use the same metric
- correctness was checked alongside speed
- the improvement or regression is stated with confidence level
- residual noise or risk is named
Output Shape
When answering with this skill, prefer:
- scenario under test
- metric and threshold
- before vs after result
- confidence level
- ship / hold recommendation
References
- [Regression checklist](./references/regression-checklist.md)
Source & license
This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.
- Author: andrewsrigom
- Source: andrewsrigom/agent-skills
- License: MIT
- Homepage: https://andrewsrigom.github.io/agent-skills/
Install and usage instructions live in the source repository linked above.
Reviews
No reviews yet, be the first.
Write a review
Versions
- v0.1.0 Imported from the upstream source.