Getting it into your agent
It runs from inside its repository, so the clone comes first — what it calls does not travel with the file alone.
git clone --depth 1 https://github.com/apisec-inc/apisec-skillsnpx agentmods add skills/apisec-inc/apisec-skills/openapi-hardenerWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/skills/apisec-inc/apisec-skills/openapi-hardener)<a href="https://agentmods.dev/skills/apisec-inc/apisec-skills/openapi-hardener"><img src="https://agentmods.dev/badge/skills/apisec-inc/apisec-skills/openapi-hardener/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/skills/apisec-inc/apisec-skills/openapi-hardener"><img src="https://agentmods.dev/badge/skills/apisec-inc/apisec-skills/openapi-hardener.svg" alt="Reviewed on agentmods" width="80" height="20"></a>What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00113 | $0.07325 |
| Opus 5.5 | $0.00045 | $0.02930 |
| Sonnet 5.5 | $0.00023 | $0.01465 |
| Haiku 4.5 | $0.00011 | $0.00732 |
Grade A, and why
openapi-hardener scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 2mo ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 1,174 lines — stays where its author put it; the contents beside it link to each section on GitHub.
OpenAPI Hardener — OWASP API3:2023
1. Role
You are an API contract security specialist who treats schema definitions as the first line of defense against bad input and data exposure. A loose schema is not a convenience — it is an attack surface. Every unconstrained string is a potential injection vector. Every missing additionalProperties: false is a mass assignment risk. Every response without an explicit field list is a data leak waiting to happen.
When reviewing or generating OpenAPI specs, JSON Schema, Zod, Joi, or Pydantic schemas, you:
- Tighten — add missing constraints (lengths, ranges, patterns, enums)
- Restrict — set
additionalProperties: false, mark server-owned fieldsreadOnly - Separate — define distinct request and response schemas (never share one schema for both)
- Produce diffs — every finding includes the exact corrected YAML or code, not general advice
2. The Security Mindset for Schemas
Input schemas: deny by default, whitelist what's allowed
An input schema defines the only fields the client is allowed to send. Everything not explicitly listed must be rejected. This is the schema equivalent of a firewall default-deny rule.
Client sends → Schema validates → Only declared fields pass through → Handler receives clean data
↓ reject
Unknown fields
Wrong types
Out-of-range values
Overlong strings
Output schemas: explicit allowlist of fields returned
An output schema defines the only fields the server will return. Without this, the serializer may pass through internal fields like passwordHash, resetToken, internalCost, or __v.
Database record → Response serializer → Only declared fields returned → Client receives safe data
↓ stripped
passwordHash
internalNotes
costPrice
__v
A loose schema is an attack surface
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 2mo ago First seen · 1,174 lines · 113 tokens per session scan A 4e5c07d17403
openapi-hardener is a skill published in the GitHub repository apisec-inc/apisec-skills (21 stars, last pushed 1mo ago), licensed MIT. It adds 113 tokens to every session and 7,325 once invoked, about $0.0005 per session on Opus 5.5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.
Other skills, from other repositories
detecting-api-enumeration-attacks
Detect and prevent API enumeration attacks including BOLA and IDOR exploitation by monitoring sequential identifier access patterns and authorization failures.
API Security Testing
Comprehensive API security testing based on OWASP API Security Top 10 including broken authentication, injection attacks, rate limiting, BOLA/BFLA vulnerabilities, and automated security scanning with ZAP and custom scripts.
audit-blocking-io-async
Blocking I/O in async contexts checklist. Use when reviewing async/await code, event-loop or reactor-based services, coroutines, or worker/thread pools — sync calls in async handlers, CPU work on the event loop, sync-over-async, blocking sleeps, missing timeouts.
audit-statelessness
Statelessness checklist for horizontally scaled apps. Use when reviewing services meant to run as multiple replicas — in-memory sessions or counters, module/static mutable state, local-disk uploads, process-local locks or schedulers, sticky-session assumptions, state lost on deploy.
audit-time-clock
Time, timezone, and clock-correctness checklist. Use when code stores or computes timestamps, durations, timeouts, or expiry — naive vs aware datetimes, server-local now(), wall-clock for elapsed time, DST and calendar math, cross-zone comparisons.
audit-api-validation
Input validation and API contract checklist. Use when reviewing request validation, DTOs, public endpoints, types, bounds, or unknown-field handling.