Agents and API
Agent skill
Let Claude Code, Codex or Cursor scan, fix and rescan for you.
The WebSecure skill teaches a coding agent the whole loop: start a scan, wait for it, read the report, fix your code, and rescan to confirm. It works with Claude Code and with any agent that can read a Markdown instruction file and run curl.
Create an API key
Create a key under API keys. Copy the key when it appears. Put it in your shell environment:
export WEBSECURE_API_KEY=ws_...
Install the skill
For Claude Code:
mkdir -p ~/.claude/skills/websecure curl -fsSL https://scan.hoopgram.com/skill/SKILL.md -o ~/.claude/skills/websecure/SKILL.md
Other agents can read the same file at /skill/SKILL.md.
Ask your agent
Open your project and ask in plain words.
Example prompts
- “Scan myapp.com with websecure and fix what it finds.”
- “Run a deep websecure scan of myapp.com. It is verified.”
- “Rescan my last websecure scan and tell me what is still open.”
- “Fix the critical and high issues from this WebSecure report.”
What the agent will and will not do
- It fixes what can be fixed in your repository, such as headers in your server or framework config, cookie flags and outdated libraries.
- It asks you before DNS, CDN, hosting or secret changes, and tells you the exact setting to change.
- It only scans sites you own or are allowed to test.
- Scans are free. For a daily check with email alerts, turn on Protect for the site ($9 a month).
No skill? Use the report prompt
Every report ends with a Copy “fix with your agent” prompt button. Paste it into any agent in your project. It contains the full report and the rescan instructions.