4 agent skills I built for the bugs AI coding agents keep shipping

Description and use

Four skills built to fix common AI agent bugs: Stripe subscription handling, data access control, tool security, and documentation accuracy.

Problem

AI coding agents frequently repeat costly mistakes such as silent Stripe failures, unauthorized data access, insecure tool configurations, and stale documentation.

Outcome

Agents consistently avoid the four highlighted bugs, improving reliability and reducing costly errors.

Workflow

• Identify recurring agent bugs
• Design a skill with rules and mappings
• Package the skill as SKILL.md
• Install the skill into the agent workflow
• Run the agent and observe corrected behavior

Evidence

Cursor, Claude Code and Codex are fast, but they keep making the same expensive mistakes.

Source

The silent Stripe bug: the webhook returns 200 but the database never updates, so cancelled users keep access, or an upgrade creates a second subscription and charges twice.

Source

The most common web flaw: user A can read user B's data.

Source

For building MCP servers, or deciding whether to install someone else's.

Source

Docs that don't lie: every claim must point to a file, command output or commit, and commands are actually run before they go in the README.

Source