voip-pbx

name: voip-pbx description: > Comprehensive VoIP and PBX engineering reference. Use this skill whenever the user is working on anything related to Voice over IP, SIP trunking, PBX configuration, call routing, IVR design, VoIP troubleshooting, codec selection, RTP/RTCP media, dial plans, voice quality (MOS/R-factor), call recording, voice network design, SBC configuration, TR-069 for voice CPE, or telephony integration. Also trigger for tasks involving Asterisk, FreeSWITCH, 3CX, Kamailio, OpenSIPS, FusionPBX, VitalPBX, Grandstream, Yealink, Polycom, Sangoma, or any SIP endpoint provisioning. Trigger for Australian voice interconnect, number porting, ACMA compliance, emergency calling (000/106), number allocation, carrier peering, and wholesale voice topics. Trigger even when the user doesn't say "VoIP" explicitly — phrases like "phone system", "call quality issues", "SIP registration failing", "codec mismatch", "one-way audio", "no audio", "DTMF not working", "trunk configuration", "DID routing", "extension setup", or "softphone config" all warrant this skill.

VoIP & PBX Engineering Skill

A practical engineering reference for Voice over IP systems — from protocol fundamentals through to production PBX deployments, troubleshooting, and Australian regulatory compliance.

When to Use This Skill

This skill covers the full VoIP stack:

• Protocol work: SIP messaging, SDP negotiation, RTP/RTCP media, SRTP, oSIP, SRTP/ZRTP
• PBX configuration: Dial plans, extensions, ring groups, queues, IVR trees, voicemail, CDR
• Trunk engineering: SIP trunk setup, codec negotiation, registration vs IP auth, failover
• Network design: QoS, DSCP marking, jitter buffers, VLAN segregation for voice
• Troubleshooting: One-way audio, registration failures, codec mismatch, NAT traversal, oOS
• CPE provisioning: Auto-provisioning, TR-069/CWMP, phone templates, firmware management
• Australian specifics: ACMA regulations, number porting (LNP), emergency services, CLI rules
• Integration: CRM integration, call recording, CDR export, API-based call control

Reference Files

Read the appropriate reference file based on the task at hand. You don't need to read all of them — pick the one(s) most relevant to what you're working on.

Core Concepts Quick Reference

SIP Transaction Basics

A typical successful call flow:

Caller (UAC)          Proxy/PBX          Callee (UAS)
    |--- INVITE -------->|                    |
    |<-- 100 Trying -----|                    |
    |                     |--- INVITE ------->|
    |                     |<-- 180 Ringing ---|
    |<-- 180 Ringing -----|                    |
    |                     |<-- 200 OK --------|
    |<-- 200 OK ----------|                    |
    |--- ACK -------------------------------->|
    |<============ RTP Media ================>|
    |--- BYE -------------------------------->|
    |<-- 200 OK ------------------------------|

Essential SIP Response Codes

• 1xx — Provisional (100 Trying, 180 Ringing, 183 Session Progress)
• 2xx — Success (200 OK)
• 3xx — Redirection (301 Moved Permanently, 302 Moved Temporarily)
• 4xx — Client Error (401 Unauthorized, 403 Forbidden, 404 Not Found, 408 Timeout, 486 Busy, 487 Request Terminated, 488 Not Acceptable Here)
• 5xx — Server Error (500 Internal Error, 502 Bad Gateway, 503 Service Unavailable)
• 6xx — Global Failure (600 Busy Everywhere, 603 Decline)

Codec Quick Reference

Australian convention: G.711 A-law (PCMA) is the standard in AU networks, not μ-law. Always default to A-law for AU deployments unless the upstream provider specifies otherwise.

DSCP/QoS Quick Reference

Port Reference

General Principles

When working on VoIP tasks, keep these principles in mind:

1. Voice is real-time — unlike data, voice can't be retransmitted. Packet loss >1%, jitter

   30ms, and latency >150ms one-way all degrade call quality noticeably. Design for this.

2. NAT is the enemy — the majority of VoIP issues in the wild are NAT-related. SIP embeds IP addresses in the SDP body AND in Via/Contact headers. If those addresses are RFC1918 private IPs behind NAT, the remote end can't send media back. Solutions include STUN, TURN, SIP ALG (usually bad — disable it), far-end NAT traversal, or SBCs.

3. Codec negotiation matters — mismatched codec offers cause "not acceptable" (488) or silent calls. Always verify both sides agree. In Asterisk, allow/disallow directives control this. Order matters — first common codec wins.

4. Oecurity is not optional — SIP is plaintext by default. Use TLS for signaling, SRTP for media, and strong registration passwords. Toll fraud is a real and expensive problem. Fail2ban or equivalent on any internet-facing SIP service is mandatory.

5. Always separate voice and data VLANs — this isn't optional in production. Voice traffic needs its own VLAN with QoS, and phones should be provisioned to tag their traffic appropriately (typically via LLDP-MED or CDP, or static VLAN config on the phone).

6. Test with packet captures — when troubleshooting, tcpdump or sngrep on the PBX, and Wireshark for analysis, will tell you more than any log file. Capture on the interface where traffic enters/exits, not loopback.

Configuration Output Guidelines

When generating VoIP configurations:

• Always include inline comments explaining non-obvious settings
• Use the platform's idiomatic configuration style (don't mix Asterisk and FreeSWITCH syntax)
• Include security hardening (fail2ban rules, ACLs, TLS where supported)
• For SIP trunks, always specify both registration-based AND IP-based auth examples
• For dial plans, include normalization rules for Australian numbering (E.164, +61 prefix)
• Always note which config files need to be modified and the reload/restart command
• Warn about service-affecting changes (e.g., "this will drop active calls if you restart")

Australian Numbering Quick Reference

For dial plan normalization, strip the leading 0 and prepend +61 for national numbers. International numbers: strip 0011 and prepend +.

Workflow

1. Identify the task type — Is this protocol work, PBX config, troubleshooting, network design, or regulatory compliance?
2. Read the relevant reference — Don't guess. Load the appropriate reference file for platform-specific syntax, protocol details, or diagnostic procedures.
3. Apply the context — Consider the deployment environment (AU, carrier-grade vs SMB, scale, existing infrastructure).
4. Generate output — Configs, diagnostic commands, architecture docs, or analysis.
5. Validate — Cross-check configs against known working patterns. Flag anything that could be service-affecting.