- Home
- Skills
- Testing & Debugging
- Run Improvement Sweeps Across Your Products
More screenshots
Works with the AI Agents you already use
Run Improvement Sweeps Across Your Products
Keep dozens of shipped products getting better: the same fixed set of real improvements for every product, tests green before anything counts as done.
$12
Run Improvement Sweeps Across Your Products
Example session with this skill installed
Initiate a sweep across the 3 repos in /internal-tools. I need a quota of 2 functional improvements each. Pilot a new Tailwind-based dark mode on 'admin-panel' before the others.
- Read your context and instructions
- Compiled the run improvement sweeps
I have completed the improvement sweep for the internal tools portfolio. One product (inventory-manager) was parked due to a failing build environment, while the other two received the full quota of functional improvements and the piloted design update.
Portfolio Sweep Report: Internal Tools
| Product | Status | Improvements | Bugs Fixed | Suite Results | Commit |
|---|---|---|---|---|---|
| admin-panel | COMPLETED | 1. Added CSV export to user logs. 2. Implemented MFA status indicators. | Fixed null pointer on empty search. | Smoke: PASS, E2E: PASS | feat: sweep-ap-772 |
| billing-gateway | COMPLETED | 1. Added Stripe webhook retry logic. 2. Implemented PDF invoice preview. | Resolved race condition in auth. | Smoke: PASS, E2E: PASS | feat: sweep-bg-910 |
| inventory-manager | PARKED | N/A | N/A | Smoke: FAIL (Node mismatch) | N/A |
Design Pilot: Tailwind Dark Mode
- Pilot Product:
admin-panel - Status: IMPLEMENTED
Changes: Replaced legacy CSS with Tailwind dark: variants. Standardized slate-900 backgrounds and cyan-500 accents.
- Rollout Status: Pending approval for
billing-gateway.
Summary Totals
- Products Processed: 3
- Functional Improvements: 4
- Bugs Fixed: 2
- Parked Products: 1
Next steps
- Review the Tailwind dark mode implementation in the
admin-panelstaging environment. - Address the Node.js version mismatch in
inventory-managerto unpark the product. - Approve the design rollout to apply the dark mode theme to
billing-gateway.
Connects securely to your tools. The creator never sees your data.
About this skill
A portfolio of dozens of shipped products decays unevenly: the ones in front of you get polished, the rest quietly rot. A sweep fixes that with quotas and gates instead of attention. In the reference run, 52 products each received 5 functional improvements — 260 in total — with every test suite green at commit, and 12 real bugs found and fixed along the way. This package is the method that produced that run.
Inside the package: the skill file with the sweep protocol and the standing rules, an example sweep report page, and a sweep checklist. The rules that carry it: the quota is the same for every product, and improvements must be functional — something a user can do or feel, never cosmetic changes filling slots. Any bug found gets fixed on sight, without asking; a sweep that logs bugs for later is a sweep that ships bugs. Smoke and end-to-end suites run before a product is declared done. Any design change or shared pattern is built on one product first, shown to the owner, and decided together before the rest are touched — and each product keeps its own look suited to its category, because a sweep never stamps one template across a portfolio.
A run goes like this: the agent inventories every product with its repo, its test command, and its current state — a product it can't build or test gets flagged on the list, not silently skipped. Then product by product: read it, pick the quota of improvements from its real gaps, build, run the full suite, commit — never batching uncommitted changes across products, so one product's breakage can't contaminate another's. A product blocked on something external is parked with its reason recorded, and the sweep continues. The report enumerates every improvement by name, per product, with bugs fixed, suite results, and commit references. Not the three highlights. The full list.
Find more skills like this on the Claude Skills Marketplace, the Codex Skills Marketplace and the Gemini Skills Marketplace, or browse Claude Skills for Testing & Debugging.
External blocks or dead dependencies cause a product to be parked rather than fixed. Functional improvements are prioritized over cosmetic-only changes unless part of an approved pilot.
How to use
- 1
Connect your agent to the Agensi MCP
Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup
- 2
Sign in to Agensi
Your agent shows a sign-in link. Sign in and click Allow.
- 3
Ask for what you need
Your agent loads this skill and uses it on your task. It always asks before buying a skill.
Reviews
No reviews yet
Be one of the first to try it. Every listed skill goes through the trust checks below.
Security scanned
Checked by our 8-point scan before listing
Fresh listing
Recently published to Agensi
Money-back guarantee
Doesn't do what its page promises? You get your money back.
Trust & safety
Security scan
Security scan passed 6 days ago
- Software
- A portfolio of reposWorking build and test commands