auto-ingest: ietf-elements backfill — 206 new elements (2026-07-12)

Author: ietf-corpus-bot

corpus/elements/rfc-9484__design-rationale-context-ids-are-split-into-even-client-allocated.yaml

@@ -0,0 +1,9 @@
+document: rfc-9484
+kind: design-rationale
+summary: Context IDs are split into even (client-allocated) and odd (proxy-allocated) values to avoid requiring synchronization between endpoints when allocating new Context IDs for protocol extensions. Once allocated, a Context ID may be used by either endpoint regardless of who allocated it.
+section: "5"
+topics:
+    - ip
+    - http
+extracted_by: claude-sonnet-4-6
+extracted_at: "2026-07-12"

corpus/elements/rfc-9484__design-rationale-the-hop-count-ttl-is-decremented-upon-encapsulation.yaml

@@ -0,0 +1,11 @@
+document: rfc-9484
+kind: design-rationale
+summary: The Hop Count (TTL) is decremented upon encapsulation rather than decapsulation, matching the behavior of IPsec. This prevents infinite forwarding loops when routing loops exist in the network. The decrement does not apply to IP packets generated by the IP proxying endpoint itself.
+section: "7.2"
+topics:
+    - ip
+    - routing
+    - vpn
+    - ipsec
+extracted_by: claude-sonnet-4-6
+extracted_at: "2026-07-12"

corpus/elements/rfc-9484__interoperability-note-icmp-traffic-is-always-allowed-through-an-ip.yaml

@@ -0,0 +1,8 @@
+document: rfc-9484
+kind: interoperability-note
+summary: ICMP traffic is always allowed through an IP proxying tunnel regardless of the 'ipproto' URI Template variable value or the IP Protocol field in ROUTE_ADVERTISEMENT entries. This ensures ICMP error signalling (e.g., Packet Too Big, Destination Unreachable) always functions.
+section: "4.6"
+topics:
+    - ip
+extracted_by: claude-sonnet-4-6
+extracted_at: "2026-07-12"

corpus/elements/rfc-9484__normative-requirement-in-route-advertisement-capsules-endpoints-must-not-send.yaml

@@ -0,0 +1,10 @@
+document: rfc-9484
+kind: normative-requirement
+summary: In ROUTE_ADVERTISEMENT capsules, endpoints MUST NOT send routes that overlap via the IP Protocol = 0 (all protocols) ambiguity. If a receiver detects such overlap, it MUST abort the IP proxying request stream (validating this is OPTIONAL, but the abort upon detection is required).
+section: 4.7.3
+rfc2119_level: MUST NOT
+topics:
+    - ip
+    - routing
+extracted_by: claude-sonnet-4-6
+extracted_at: "2026-07-12"

corpus/elements/rfc-9484__normative-requirement-ip-proxies-must-validate-whether-the-decoded-target.yaml

@@ -0,0 +1,11 @@
+document: rfc-9484
+kind: normative-requirement
+summary: IP proxies MUST validate whether the decoded 'target' and 'ipproto' URI Template variables meet format requirements. If 'target' is a DNS name, the IP proxy MUST perform DNS resolution before replying. On DNS errors, the proxy MUST reject the request and SHOULD send a Proxy-Status header field with the dns_error proxy error type.
+section: "4.1"
+rfc2119_level: MUST
+topics:
+    - http
+    - ip
+    - dns
+extracted_by: claude-sonnet-4-6
+extracted_at: "2026-07-12"

corpus/elements/rfc-9484__normative-requirement-ip-proxying-endpoints-that-operate-as-routers-and.yaml

@@ -0,0 +1,10 @@
+document: rfc-9484
+kind: normative-requirement
+summary: IP proxying endpoints that operate as routers and support IPv6 MUST ensure that the IP tunnel link MTU is at least 1280 bytes. If using QUIC DATAGRAM frames and the QUIC MTU is too low to allow sending 1280 bytes, the endpoint MUST abort the IP proxying request stream.
+section: "7.2"
+rfc2119_level: MUST
+topics:
+    - ip
+    - quic
+extracted_by: claude-sonnet-4-6
+extracted_at: "2026-07-12"

corpus/elements/rfc-9484__normative-requirement-ip-proxying-over-http-must-be-operated-over.yaml

@@ -0,0 +1,12 @@
+document: rfc-9484
+kind: normative-requirement
+summary: IP proxying over HTTP MUST be operated over TLS or QUIC encryption, or another equivalent encryption protocol, to provide confidentiality, integrity, and authentication.
+section: "4"
+rfc2119_level: MUST
+topics:
+    - http
+    - ip
+    - security
+    - crypto
+extracted_by: claude-sonnet-4-6
+extracted_at: "2026-07-12"

corpus/elements/rfc-9484__normative-requirement-the-uri-template-used-to-configure-clients-must.yaml

@@ -0,0 +1,10 @@
+document: rfc-9484
+kind: normative-requirement
+summary: The URI Template used to configure clients MUST be a level 3 template or lower, MUST be in absolute form with non-empty scheme, authority, and path components, and MUST NOT contain non-ASCII Unicode characters or use Reserved/Fragment/Label/Path-Segment expansion operators. If a client detects any requirement violation, it MUST reject its configuration and abort the request.
+section: "3"
+rfc2119_level: MUST
+topics:
+    - http
+    - ip
+extracted_by: claude-sonnet-4-6
+extracted_at: "2026-07-12"

corpus/elements/rfc-9484__normative-requirement-when-implementing-scoping-or-routing-by-internet-protocol.yaml

@@ -0,0 +1,9 @@
+document: rfc-9484
+kind: normative-requirement
+summary: When implementing scoping or routing by Internet Protocol Number, implementations MUST walk the chain of IPv6 extension headers to find the outermost non-extension Internet Protocol Number when matching against scoping rules. IP proxies MAY reject requests scoped to extension header protocol numbers.
+section: "4.8"
+rfc2119_level: MUST
+topics:
+    - ip
+extracted_by: claude-sonnet-4-6
+extracted_at: "2026-07-12"

corpus/elements/rfc-9484__normative-requirement-when-ip-packet-forwarding-fails-e-g-no.yaml

@@ -0,0 +1,9 @@
+document: rfc-9484
+kind: normative-requirement
+summary: When IP packet forwarding fails (e.g., no route, policy rejection, MTU exceeded), IP proxying endpoints SHOULD use ICMP/ICMPv6 to signal the forwarding error to the peer by generating ICMP packets and sending them via HTTP Datagrams. Relevant errors include Destination Unreachable and Packet Too Big.
+section: 7.2.1
+rfc2119_level: SHOULD
+topics:
+    - ip
+extracted_by: claude-sonnet-4-6
+extracted_at: "2026-07-12"