- Home
- Skills
- Agents & Orchestration
- Showwork
Showwork
Explain completed software changes with visuals, checks and remaining problems.
Free
Works with the AI Agents you already use
Showwork
Example session with this skill installed
/showwork explain the changes I just made to the login flow and show me if the validation still works.
- Read your context and instructions
- Compiled the showwork
I have analyzed the changes to the login flow. The implementation is complete, and the validation logic has been verified against the new requirements.
Login Flow Result Walkthrough
| Category | Details |
|---|---|
| Primary Change | Updated auth-service.ts to support MFA tokens and revamped the LoginForm UI. |
| Before/After | Previously, invalid emails failed silently. Now, a Zod-validated error message appears instantly. |
| Key Files | src/components/Login.tsx (UI), src/lib/validation.ts (Rules), src/api/auth.ts (API). |
| Verification | Passed: 3 unit tests, 1 manual UI render check. Warning: SMS gateway mocked. |
| Remaining | The "Forgot Password" link is currently unstyled and needs a separate ticket. |
Next steps
- Open the Showwork App to view the recorded screenshot of the new validation error state.
- Review the
Architecture/Decisions/backendnote I saved to your Obsidian vault. - Run
npm testto confirm the full suite passes in your local environment.
Connects securely to your tools. The creator never sees your data.
About this skill
Stop guessing what your coding agent actually accomplished. Coding agents often leave you with a "done" message and a mess of files to sort through. Showwork provides a visual walkthrough of software changes, allowing you to see exactly what changed, how it works now, and what remains unresolved.
What it does
- Visual Walkthroughs generates a browser-based report showing screens, before/after behavior, and critical file changes.
- Implementation Planning investigates code to create testable, visual plans with concrete success criteria before writing a single line of code.
- Adversarial Reviews analyzes coding sessions, tool results, and diffs to identify risks and unverified claims.
- Behavior Verification runs focused checks, runtime exercises, and UI inspections to prove requirements are met.
- Decision Tracking saves architectural choices and system changes directly to an Obsidian vault for long-term documentation.
How it works
- Invoke Commands use
/showworkto explain a result,/showwork-planfor research, or/showwork-reviewfor a deep audit. - Launch Companion triggers a local Python companion that opens the Showwork macOS app to host the report.
- Record Evidence captures screenshots, test logs, and API results using the internal evidence recorder.
- Publish Report delivers a final interactive walkthrough containing the diff, visual proof, and remaining issues.
Frameworks & tools
Requires Node.js 20+, Python 3.11+, and Git. Integrates with the Showwork macOS app for visual reports and Obsidian for architectural decision records.
Why this beats prompting it yourself
Generic prompts often result in a list of files without context. Showwork uses a dedicated local companion and evidence recorder to verify runtime behavior and UI changes, ensuring the explanation is based on reality rather than model hallucinations.
Use cases
- Post-task Audits verify that an agent actually fixed a bug instead of just moving code around.
- Visual UI Reviews see rendered web pages or iOS Simulator screens inside a dedicated report tab.
- Architecture Logging automatically sync project decisions to Obsidian to maintain a "source of truth."
- Complex Planning resolve implementation choices and dependencies before starting large refactors.
Known limitations
Requires a macOS companion app for visual reports. Cannot modify production code during review or planning phases unless explicitly authorized.
Requires macOS app for visual reports. Needs Node.js 20+ and Python 3.11+ locally. Limited to projects with accessible local Git history.
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 any payment.
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
Free forever
No account required to browse
Trust & safety
Security scan
Security scan passed 2 days ago
- Free with an Agensi account
Requires