- Home
- Skills
- DevOps & Cloud
- deploydoctor
deploydoctor
Evidence-based deployment diagnostics for Next.js, Docker, and Nginx with automated secret redaction and verified fixes.
$4.99
Works with the AI Agents you already use
deploydoctor
Example session with this skill installed
My Next.js app is failing to deploy on the VPS. Nginx is giving a 502 error. I have shell access. Can you figure out why?
- Read your context and instructions
- Compiled the deploydoctor
I have analyzed the system using scripts/collect-diagnostics.sh and reviewed the redacted logs from Nginx and the application service.
Diagnosis Report: Next.js Upstream Failure
| Metric | Status |
|---|---|
| Failure Class | Proxy / Start-up |
| Confidence | High |
| Evidence Tag | OBSERVED: Nginx logs show connect() failed (111: Connection refused) while connecting to 127.0.0.1:3000. |
| Evidence Tag | OBSERVED: `ss -ltnp |
| Evidence Tag | INFERRED: The Next.js process is either failing to start or binding to a different port/interface. |
Ranked Hypotheses
Missing Environment Variable (Confidence: High): The application requires DATABASE_URL to boot, but scripts/collect-diagnostics.sh --check-env-var DATABASE_URL returned unset.
2.
Port Conflict (Confidence: Low): Another process is preventing the bind, though no process was observed on port 3000.
Recommended Fix
Add the missing DATABASE_URL to the .env.production file or the systemd service environment block.
# Proposed Action: Verify environment variable presence
export $(cat .env.production | xargs) && npm run start
Rollback: Remove the added environment variable from the config and restart the service.
Next steps
- Approve the command to check if the app starts manually with the variable set.
- Provide the correct database connection string (do not paste it here; I will prompt for the file location).
- Run the verification command to confirm the 502 error is resolved.
Connects securely to your tools. The creator never sees your data.
About this skill
Deployment failures often lead to a cycle of guessing, patching, and failing again. DeployDoctor breaks this loop by enforcing a evidence-based diagnostic workflow that distinguishes between observed facts and inferred guesses. It prioritizes read-only investigation to ensure that fixes are based on proven root causes rather than plausible theories.
What it does
- Classifies failures by mapping symptoms to specific categories like build-time, startup crashes, proxy errors, or resource exhaustion.
- Ranks hypotheses by confidence levels and identifies discriminating tests to prove or disprove potential root causes.
- Redacts secrets automatically by piping logs through a specialized script to ensure no credentials ever appear in reports.
- Proposes minimal fixes that target the verified issue with the smallest possible change and a defined rollback plan.
- Verifies results using actual command output to confirm the specific symptom is gone, refusing to claim a fix based on assumptions.
How it works
- System Baseline: Establish the runtime, environment, and recent changes to define what "working" looks like.
- Evidence Collection: Run diagnostic scripts or tail logs from Docker, Nginx, or systemd, ensuring all output is redacted.
- Hypothesis Testing: Separate facts from guesses, ranking causes by confidence and running tests to isolate the true fault.
- Controlled Remediation: Present the smallest safe fix for approval before execution, following a strict safety policy.
Frameworks & tools
This skill supports Next.js, Node.js, Rust (Actix), Docker, Nginx, PostgreSQL, and GitHub Actions. It utilizes Bash, Python 3.8+, and POSIX shell utilities for log analysis and data collection.
Why this beats prompting it yourself
Generic prompts often jump to "plausible" solutions that result in configuration drift or security regressions. This skill enforces a "read-only first" discipline and an evidence-tagging system (OBSERVED, INFERRED, UNKNOWN) that prevents the agent from hallucinating causes or suggesting dangerous fixes like chmod 777.
Use cases
- Diagnose why a container is crash-looping or exiting with specific codes like 137.
- Resolve Nginx 502/503/504 errors by correlating proxy logs with upstream app behavior.
- Fix failed database migrations or connection timeouts during new deployments.
- Debug CI/CD pipeline failures where the build passes locally but fails in production.
Known limitations
Requires log access or shell visibility to be effective. It does not support greenfield architecture design or general feature development. It will refuse to disable security controls as a fix.
Find more skills like this on the Claude Skills Marketplace, the Codex Skills Marketplace and the Gemini Skills Marketplace, or browse Claude Skills for DevOps & Cloud.
Requires logs or shell access for high-confidence results. Cannot perform feature development or total rewrites. Refuses to disable security controls (CORS, TLS, Firewalls) as a solution.
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 16 days ago
- Software
Python
Node.jstsc
Git
Permissions
Allowed Hosts