mirror of
https://github.com/CyberSecurityUP/NeuroSploit.git
synced 2026-09-29 12:31:43 +02:00
CyberSecurityUPandClaude Opus 5
8894649ccb
Hard scoping was already enforced in code (every request passes
ScopePolicy::check_request; exclude beats allowlist; capability token caps
it; out-of-scope findings withheld + audited). What was missing was a way to
author that boundary from a file or the web form instead of only CLI flags.
- scope.rs: ScopePolicy::from_yaml / from_file — a dependency-free parser for
the friendly string format (app.example.com, *.wildcard, CIDR, url-prefix),
the same strings Pattern::parse already takes, NOT the raw serde {kind,value}
shape. Strict in one direction: an unreadable file errors, an empty hard list
authorizes nothing (a safe failure, but the operator's choice, not a typo).
- CLI: --scope-file <yaml>. Loaded before authorization so --in-scope adds to
it and the capability grant still caps it.
- Web: a full Scoping & Guardrails section in the Authorization tab — hard
scope, exclusions, observe-only, destructive-method + account-creation
toggles, max accounts, rate limit, forbidden payloads, notes. The server
materializes a scope YAML and passes --scope-file; notes stay labelled
"guidance, NOT enforced" so prose is never mistaken for a control.
- examples/scope.example.yaml documents the format.
End-to-end verified: web form -> YAML -> Rust loader -> enforced boundary.
332 tests (+4).
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
68 lines
2.9 KiB
YAML
68 lines
2.9 KiB
YAML
# NeuroSploit — engagement scope & guardrails
|
||
# ---------------------------------------------------------------------------
|
||
# HARD scope is enforced in code: a request whose host is not covered by `hard`
|
||
# (or is hit by `exclude`) is REFUSED before it leaves — not warned about,
|
||
# refused. SOFT scope is the guardrails inside that boundary.
|
||
#
|
||
# Every pattern below is a plain string, parsed the same way the --in-scope flag
|
||
# and the web form parse it (crate::scope::Pattern::parse):
|
||
#
|
||
# app.example.com exact host
|
||
# *.example.com the apex AND every sub-domain
|
||
# 10.0.0.0/24 an IPv4 network (CIDR)
|
||
# https://example.com/api/v2 a URL prefix — narrower than a whole host
|
||
#
|
||
# An empty `hard` list means NOTHING is authorized. Scope is never implicit.
|
||
# ===========================================================================
|
||
|
||
# --- HARD: the allowlist. Only these are testable. ------------------------
|
||
hard:
|
||
- app.example.com
|
||
- "*.staging.example.com" # apex + subdomains of the staging tier
|
||
- https://example.com/api/v2 # only this path prefix on the apex host
|
||
- 10.20.30.0/24 # an internal range reached via --transport
|
||
|
||
# --- EXCLUDE: carve-outs. These always beat the allowlist. ----------------
|
||
# A host here is refused even if `hard` would otherwise cover it.
|
||
exclude:
|
||
- admin.example.com # never touch the admin console
|
||
- https://app.example.com/billing # PCI surface — out of this engagement
|
||
- payments.example.com
|
||
|
||
# --- SOFT: guardrails inside the boundary ---------------------------------
|
||
soft:
|
||
# Hosts you may LOOK at but never attack (recon only — no payloads).
|
||
observe_only:
|
||
- cdn.example.com
|
||
- "*.thirdparty.example.com"
|
||
|
||
# State-mutating verbs (DELETE/PUT/PATCH). Off by default: a scan should not
|
||
# change the target's state to "prove" a bug.
|
||
allow_destructive_methods: false
|
||
|
||
# Registering test accounts, and how many. 0 = unlimited (not recommended).
|
||
allow_account_creation: true
|
||
max_accounts: 3
|
||
|
||
# Requests per minute across the WHOLE engagement. 0 = unlimited.
|
||
# Keep this low on production; the OT profile caps far lower still.
|
||
max_requests_per_minute: 240
|
||
|
||
# Payload substrings that are NEVER acceptable, whatever the finding — the
|
||
# classes that damage a production target instead of demonstrating a bug.
|
||
# These extend the built-in defaults (drop table, rm -rf /, fork bombs, …).
|
||
forbidden_payloads:
|
||
- "drop table"
|
||
- "truncate table"
|
||
- "delete from"
|
||
- "rm -rf /"
|
||
- "shutdown"
|
||
- "while(true)"
|
||
|
||
# Free-text context for the agents. NOT enforceable — kept separate from the
|
||
# rules on purpose, so nobody mistakes prose for a control.
|
||
notes:
|
||
- "Authorized per SOW-2026-0142; contact security@example.com on any outage."
|
||
- "Test window 02:00–06:00 UTC only."
|
||
- "Data-exfil PoCs: prove read access with a canary row, do not pull real PII."
|