More screenshots
Works with the AI tools you already use
Claude Code
Cursor
Codex CLI
GitHub Copilot
Gemini CLI+20 moreClaude CodeCursorCodex CLIGitHub CopilotGemini CLI
OpenClaw
Manus
Grok+17 agentsAI Automation Incident Intake Assistant
Standardize messy AI automation failure reports into evidence-based incident intake records.
Free
AI Automation Incident Intake Assistant
Example session with this skill installed
Client says the OpenAI bot stopped working at 2pm. They think it's the API key. I saw a 401 error in the Make logs for one execution. They want it fixed now for a launch tomorrow. The developer says the credentials were rotated yesterday and added to the workflow, but nobody has actually tested them.
- Read your context and instructions
- Compiled the ai automation incident
Incident Intake Summary
A customer-facing AI automation is reported to have stopped working at approximately 14:00. A 401 Unauthorized error was reportedly observed in one Make execution. The client suspects an API-key problem, while the developer reports that credentials were rotated and added but not tested.
Evidence Status
| Item | State | Establishes | Does NOT Establish |
|---|---|---|---|
| Bot stopped working | OBSERVED | Customer-facing failure is reported | Exact failure point or scope |
| 401 in Make logs | OBSERVED | A 401 was reportedly seen in one execution | Current failure or root cause |
| Credentials rotated | OBSERVED | Credential rotation occurred | Credentials are valid |
| Credentials added | OBSERVED | Configuration activity occurred | Successful authentication |
| Credentials tested | UNKNOWN | No successful test supplied | Whether authentication works |
Severity
USER-PROVIDED SEVERITY: NOT PROVIDED
EVIDENCE-SUPPORTED SEVERITY: UNKNOWN — INSUFFICIENT EVIDENCE
A customer-facing failure is reported, but the supplied information does not establish its scope, duration, number of affected customers, or consequential business impact sufficiently to assign a supported severity level.
URGENCY: HIGH
The launch deadline supports rapid investigation but does not independently establish P0 or P1 severity.
Root Cause
NOT DETERMINED
Authentication/credential failure is a strong investigation area because a 401 is reportedly present.
It remains a hypothesis until the relevant execution evidence and authentication behavior are verified.
Other possible failure points remain open.
Safe to Resume
NOT ESTABLISHED
No evidence demonstrates successful authentication, successful workflow execution, successful AI processing, or successful downstream delivery.
Highest-Value Evidence Needed
- One sanitized affected execution.
- The corresponding sanitized 401 response/log.
- Whether the affected execution reached the API authentication step.
- An authorized authentication test result.
- Evidence showing whether the failure is ongoing.
Do not provide passwords, API keys, access tokens, or other secrets.
Handoff
Known: customer-facing failure is reported and a 401 is reportedly present.
Suspected: authentication/credential failure.
Not established: root cause, credential validity, failure location, scope, recovery, and safe resumption.
Next action: collect one affected execution trace and the corresponding sanitized API response before assigning causation or declaring the automation safe to resume.
Connects securely to your tools. The creator never sees your data.
About this skill
The problem
Vague reports like “something is broken” leave automation builders guessing. Missing evidence, contradictory reports, and unverified root-cause claims can lead to wasted investigation time and unsafe production decisions.
What it does
- Standardizes messy AI automation failure reports into structured, evidence-based incident intake records.
- Separates reported facts, observations, hypotheses, unknowns, and potential risks.
- Preserves contradictory client, developer, and workflow-owner reports without choosing sides.
- Prevents suspected root causes from being presented as confirmed causes.
- Separates severity from urgency so deadlines and executive pressure do not automatically inflate technical severity.
- Identifies what is actually evidenced and what still requires verification.
- Creates targeted evidence-collection requests and an actionable investigation handoff.
- Tracks reproduction, recovery, and safe-to-resume status without inventing completed actions.
- Handles AI-specific failure chains from AI input and output through workflow decisions and downstream business actions.
- Redacts exposed secrets instead of requesting, storing, or reproducing credentials and API keys.
Evidence-first incident handling
The skill uses explicit evidence states such as VERIFIED, OBSERVED, INFERRED, POTENTIAL, UNKNOWN, NOT CHECKED, and NOT APPLICABLE.
It does not claim live-system inspection, reproduction, recovery, successful authentication, or safe resumption unless the supplied evidence supports the claim.
Severity discipline
User-provided severity is preserved separately from evidence-supported severity. A P0 request, launch deadline, executive concern, customer-facing workflow, or potentially serious consequence does not automatically establish P0 or P1 severity.
When the available evidence is insufficient to establish consequential impact, the skill uses:
SEVERITY: UNKNOWN — INSUFFICIENT EVIDENCE
Potentially critical consequences remain recorded as potential risks rather than being converted into confirmed severity.
Useful for
- n8n automations
- Make workflows
- Zapier automations
- AI agents
- OpenAI and other LLM integrations
- APIs and webhooks
- CRMs and databases
- WhatsApp and messaging automations
- Customer-facing business workflows
- Automation agencies and freelance builders
Compatibility
Designed as a reusable SKILL.md-based workflow for compatible AI coding agents and AI assistants that support custom skills/instructions, including Claude, Cursor, Codex CLI, GitHub Copilot, Gemini, and OpenClaw. Exact skill-loading behavior depends on the host platform.
Important limitation
This is an incident intake and evidence-structuring skill, not a live monitoring, debugging, deployment, or incident-response platform. It analyzes the information and evidence you provide. It does not directly access client systems, execute workflows, modify automations, send messages, or store credentials.
How to install
Works the same in every agent - Claude, Cursor, Codex, Copilot and 20+ more.
- 1
Download the ZIP
Free skills download straight away. Paid skills unlock right after purchase.
- 2
Unzip into your skills folder
Every agent reads skills from one folder on your machine. Drop the unzipped folder in there.
- 3
Ask your agent to use it
Restart the agent if it was already running. It picks the skill up automatically - no config needed.
Skills folder by agent
Click the path to copy it. Create the folder if it does not exist yet.
Reviews
No reviews yet
Be one of the first to try it. Every listed skill passes our trust checks below.
Security scanned
Passed our 8-point scan before listing
1 install
Downloaded by developers to date
Free forever
No account required to browse
Trust & safety
Security scanned
Verified clean 7 days ago
- Free to download with an account