Guide · Agent skill
How to use the API Security Assessment agent skill (api-security)
Updated · written with AI from the project's own documentation and checked against our catalogue data
The api-security skill is a reusable instruction set that gives an AI agent a structured 10-phase playbook for assessing web APIs (REST, GraphQL, WebSocket and SOAP) that you are explicitly authorized to test. This guide explains what the skill changes about the agent's behaviour, how to install it, how to trigger it, and what safeguards to expect. By the end you will know whether it fits your work and how to use it without crossing scope.
API Security Assessmentsickn33/agentic-awesome-skills · MIT Download the skill Quick start
- 1Get the ZIP. Download the api-security skill ZIP from this page.
- 2Unzip into a skills folder. Unzip into ~/.claude/skills/ for a personal install, or into a project's .claude/skills/ for a project install.
- 3Or upload it in the Claude apps. Go to Settings → Capabilities → Skills and upload the skill.
- 4State the target and authorization. In your request, name the exact target and confirm you have written permission and the permitted scope.
- 5Review the plan. Ask the agent for the planned checks and the exact commands before anything intrusive runs.
- 6Approve, run and record. Confirm each command you accept, then keep reproducible evidence and a list of checks not executed.
What the api-security skill makes the agent do
An agent skill is a folder with a SKILL.md file and optional reference documents that the agent reads when a task matches the skill's description. This one turns the agent into a guided assistant for API security assessments. It covers endpoint discovery, JWT and OAuth login weaknesses, broken object-level and function-level authorization (BOLA/IDOR and BFLA), input handling, rate limiting, WebSocket and GraphQL specifics, data over-exposure, and reporting.
The skill is offensive in nature, so its most important behaviour is the confirmation gate. Before running any command that probes, exploits, changes data, or attempts credential access, the agent must ask for the exact target, confirm written authorization and scope, show the exact commands, and wait for your explicit confirmation. Without that, it stays read-only and gives defensive guidance.
- Phase 1: API discovery and reconnaissance
- Phase 2: Authentication testing (JWT, OAuth 2.0)
- Phase 3: Authorization testing (BOLA/IDOR, BFLA)
- Phase 4: GraphQL-specific checks
- Phase 5: REST input validation
- Phase 6: Business logic and differential testing
- Phase 7: WebSocket testing
- Phase 8: Rate limiting and DoS
- Phase 9: Data exposure
- Phase 10: CI/CD integration
Install the api-security skill
Download the ZIP for the skill from this page and unzip it. You can install it in one of two places. For a personal install available in all projects, unzip it into ~/.claude/skills/. For a single project, unzip it into that project's .claude/skills/ folder so it travels with the repository.
If you use the Claude apps, you can instead upload the skill under Settings → Capabilities → Skills.
- Personal install: ~/.claude/skills/ (available across projects)
- Project install: .claude/skills/ inside your project folder
- Claude apps: Settings → Capabilities → Skills, then upload the skill
How to trigger the skill with natural requests
You do not need a special command. Describe the task in plain language and name the target and your authorization. The agent matches the request to the skill's description, which covers REST, GraphQL, WebSocket and SOAP APIs, authentication and authorization flaws, rate-limit testing and the methodology.
Include the target and scope up front. If you leave them out, the agent will ask for them before any intrusive step.
- "I have written authorization to test the staging API at the URL I'll give you. Walk me through the discovery phase."
- "Review my JWT handling for this API token and tell me which checks you would run, but do not send requests yet."
- "Give me a BOLA/IDOR checklist for our orders endpoints that I can use in our self-audit before release."
- "Draft the report structure for our API security review, based on the 10-phase methodology."
What the skill needs before you start
The skill requires explicit written authorization and a defined scope from the system owner. Agree request-rate and concurrency budgets and stop conditions before testing, and prefer a sandbox, disposable VM or controlled lab. Use controlled canary objects rather than deleting real users.
The skill names several external tools, such as jwt_tool, Burp Suite with the Autorize extension, Vespasian, and Entropy. These are optional references. They are not bundled with the skill and the skill does not verify they are installed. The tool-index and RULES files it mentions are not included.
- Written authorization and defined scope from the system owner
- Agreed rate limits, concurrency budgets and stop conditions
- Optional external tools, installed separately if you choose to run them
- A sandbox or non-production mirror where possible
Safety notes: scripts and network access
This skill does not run scripts itself, but it does use the network. Any test you approve will send real requests to the target, and some checks are intrusive, including denial-of-service style tests, Slowloris, and credential cracking. Those require specific scope and should not run against production systems without agreed limits.
Some findings are not proof on their own. Introspection being enabled or a GraphQL IDE being exposed is not, by itself, a vulnerability. Treat the output as a lead to verify, and record reproducible evidence for each finding.
- No scripts are bundled or run by the skill itself
- Network use is real: approved commands send live traffic to the target
- Intrusive checks (DoS, Slowloris, credential cracking, metadata access, persistent CI scanning) need explicit scope
- Do not claim introspection or an exposed IDE alone as a vulnerability
Tips for getting useful results
Give the agent the API's documentation or spec if you have one, such as an OpenAPI file, and two test accounts with different roles so authorization checks can compare responses. Ask it to show each command before running it, and to mark which checks were not executed, so your report is honest about coverage.
Keep your own checklist in sync with the phases. The skill is a methodology, so you still decide which checks fit your scope.
- Provide an API spec and at least two accounts with different roles
- Ask for a plan first, then approve commands one by one
- Ask the agent to list checks it did not run
- Use known, authorized object IDs rather than trying to guess UUIDs
What people use it for
- Pre-release self-audit of your own API's authentication and authorization
- Structured discovery and checklist for a scheduled penetration test you are authorized to run
- Checking whether one user can read another user's objects by changing an ID (BOLA/IDOR) in a staging environment
- Reviewing JWT and OAuth login handling against a known checklist
- Drafting an API security review report that follows a consistent methodology
Frequently asked questions
Is the api-security skill free?
The skill is published under the MIT licence, as stated in its metadata. The external tools it names, such as jwt_tool and Burp Suite, have their own licences and are not bundled with the skill.
How do I install the api-security skill?
Download the ZIP from this page and unzip it into ~/.claude/skills/ for a personal install or into a project's .claude/skills/ folder. In the Claude apps, you can upload it under Settings → Capabilities → Skills.
What do I need to use the api-security skill?
You need explicit written authorization and a defined scope from the system owner. Optional external tools are needed only if you choose to run them, and they must be installed separately.
Does the api-security skill include a scanner?
No. The skill is a methodology and checklist. The tools it names are references and are not bundled or verified as installed, so it is not a turnkey scanner.
Can I use the api-security skill on any API?
No. Only test APIs you own or have written permission to assess. The skill itself is offensive, and misuse is illegal. Use a sandbox or controlled lab where possible.
Is it safe to run against production?
Only with agreed rate limits, concurrency budgets and stop conditions. Many checks, including DoS, Slowloris and credential cracking, are intrusive, so non-production mirrors are preferred.
New agent skills, repos and AI tools — every week
The best new skills and open-source projects, the fastest-growing repos, and one free prompt. Five minutes to read. Unsubscribe anytime.
Our verdict
Choose this skill if you are a security engineer or pentester with written authorization for a specific API, or a developer who wants a structured checklist to self-audit authentication and authorization before release. Pick something else if you want a turnkey scanner, or if you lack written permission or agreed stop conditions for the target.
Download the skill