mcaxl/tests/test_prompts_package.py
Ryan Malloy 91bd3a0705 prompts: add failed_fax_investigation (cross-server, story D)
Third cross-server prompt landing. Story D from
axl/agent-threads/cross-server-prompts/. Lives in mcaxl because the
dial plan is the first layer to investigate — "did the call make it
past CUCM at all?" determines the rest of the triage.

Composes:
  - mcaxl (required)     — route lookup for the dialed DID, identify
                            egress trunk and any translation patterns
  - mcsiphon (recommended) — CDR record(s) for the failure window;
                              extract Q.850 cause codes, duration,
                              connect status, release direction,
                              device names + IPs
  - mcdewey (optional)   — Cisco's published troubleshooting guidance
                            for the specific cause code surfaced

Verdict layer names (literal vocabulary so downstream tooling can
pattern-match):
  - cucm_dial_plan
  - cucm_region_or_css
  - cube_sip_trunk_negotiation
  - far_end_sbc
  - far_end_fax_server
  - t38_negotiation_failure
  - inconclusive

Cause-code-to-layer mapping table embedded in the prompt body covers
the high-frequency Q.850 codes (1, 16, 17, 19, 27, 31, 38, 47, 65,
79, 127). Plus releaseDirection (originator/destination/network)
guidance for narrowing.

Common patterns surfaced explicitly:
  - Setup failures with dateTimeConnect=0 (signaling-layer)
  - Mid-call drops with non-zero duration (often T.38 renegotiation)
  - Cause-code mismatch between origCause and destCause (CUBE
    translation layer)
  - Receive-and-abandon at the route point (CSS doesn't reach
    destination partition)

The prompt's specific high-leverage detail is the "T.38 renegotiation
mid-call" pattern — calls that connect on G.711, then fail when fax
tones trigger T.38 switchover. That shape is invisible from any single
MCP and is exactly what cucx-docs's planned-but-unwritten
runbooks/rightfax-failed-fax-investigation.mdx anticipated.

Tests: registration sentinel updated to 14 prompts. 238 tests passing.

Closes the planned cucx-docs runbook page — that becomes a thin
operational shim around this prompt rather than a from-scratch
troubleshooting tree.
2026-05-05 20:39:42 -06:00

296 lines
10 KiB
Python

"""Tests for the extracted prompts package.
The render() functions are pure (they take docs as a parameter, no module
globals), so they're trivially unit-testable. We verify each one renders
without raising both with and without a docs index.
"""
import json
from pathlib import Path
import pytest
from mcaxl import prompts
from mcaxl.docs_loader import DocsIndex
@pytest.fixture
def fake_docs(tmp_path: Path) -> DocsIndex:
chunks = [
{
"id": "cucm::v15::admin::Route-Plan::0",
"text": "Route plan defines call routing.",
"heading_path": ["Route Plan"],
"source_path": str(tmp_path / "fake.md"),
"product": "cucm",
"version": "v15",
"doc": "system-config-guide",
},
{
"id": "cucm::v15::admin::SIP-Trunk::0",
"text": "SIP trunks connect CUCM to other call control entities.",
"heading_path": ["SIP Trunk Configuration"],
"source_path": str(tmp_path / "fake.md"),
"product": "cucm",
"version": "v15",
"doc": "system-config-guide",
},
]
(tmp_path / "chunks.jsonl").write_text(
"\n".join(json.dumps(c) for c in chunks)
)
(tmp_path / "index_meta.json").write_text(
json.dumps({"model_name": "test", "embedding_dim": 384, "products": ["cucm"]})
)
idx = DocsIndex.load(tmp_path)
assert idx is not None
return idx
# ---- route_plan_overview ----------------------------------------------------
def test_route_plan_overview_renders_with_docs(fake_docs):
text = prompts.route_plan_overview.render(fake_docs)
assert "Route Plan Overview" in text
assert "axl_version" in text
assert "route_partitions" in text
def test_route_plan_overview_renders_without_docs():
# Graceful degradation — no docs index should still produce a usable prompt
text = prompts.route_plan_overview.render(None)
assert "Route Plan Overview" in text
assert "cisco-docs index is not loaded" in text
# ---- investigate_pattern ---------------------------------------------------
def test_investigate_pattern_includes_pattern_in_output(fake_docs):
text = prompts.investigate_pattern.render(fake_docs, "10.911", "CER911-PT")
assert "10.911" in text
assert "CER911-PT" in text
def test_investigate_pattern_no_partition(fake_docs):
text = prompts.investigate_pattern.render(fake_docs, "9.@")
assert "9.@" in text
# The partition clause should not appear when no partition is given
assert "in partition" not in text
# ---- audit_routing ----------------------------------------------------------
def test_audit_routing_default_focus(fake_docs):
text = prompts.audit_routing.render(fake_docs)
assert "focus: `full`" in text
def test_audit_routing_custom_focus(fake_docs):
text = prompts.audit_routing.render(fake_docs, "translations")
assert "focus: `translations`" in text
# ---- cucm_sql_help ----------------------------------------------------------
def test_cucm_sql_help_includes_question(fake_docs):
q = "How do I find phones with no associated user?"
text = prompts.cucm_sql_help.render(fake_docs, q)
assert q in text
assert "axl_describe_table" in text
def test_cucm_sql_help_handles_empty_question(fake_docs):
# No substantive keywords — must still render something useful
text = prompts.cucm_sql_help.render(fake_docs, "x y z")
assert "axl_list_tables" in text
# ---- sip_trunk_report (NEW) ------------------------------------------------
def test_sip_trunk_report_includes_query_1(fake_docs):
text = prompts.sip_trunk_report.render(fake_docs)
# Query 1 essentials: device JOIN sipdevice, the trunk_name column,
# and the WHERE filter on Trunk class
assert "FROM device d" in text
assert "JOIN sipdevice sd" in text
assert "tc.name = 'Trunk'" in text
def test_sip_trunk_report_includes_query_2_and_tools(fake_docs):
text = prompts.sip_trunk_report.render(fake_docs)
assert "siptrunkdestination" in text
# And references the existing MCP tool for route group traversal
assert "route_lists_and_groups()" in text
def test_sip_trunk_report_with_name_filter_injects_safe_like(fake_docs):
text = prompts.sip_trunk_report.render(fake_docs, "PSTN")
# The filter should appear as a LIKE clause and the scope note should
# reflect the narrowing
assert "%PSTN%" in text
assert "narrowed to trunks" in text
def test_sip_trunk_report_name_filter_escaped(fake_docs):
# SQL injection defense: a single-quote in the name_filter must be
# doubled before going into the LIKE pattern
text = prompts.sip_trunk_report.render(fake_docs, "O'Reilly")
assert "O''Reilly" in text
def test_sip_trunk_report_renders_without_docs():
text = prompts.sip_trunk_report.render(None)
assert "SIP Trunk Report" in text
assert "cisco-docs index is not loaded" in text
# ---- registration smoke test -----------------------------------------------
def test_all_prompts_registered_in_server():
"""Confirm each prompt module's render() is wired through a FastMCP shim
in server.py. Catches the case where a new module is added but the
shim wasn't (the prompt would be invisible to the LLM)."""
import asyncio
from mcaxl import server
async def _list():
registered = await server.mcp.list_prompts()
return {p.name for p in registered}
names = asyncio.run(_list())
assert names == {
"route_plan_overview",
"investigate_pattern",
"audit_routing",
"cucm_sql_help",
"sip_trunk_report",
"phone_inventory_report",
"user_audit",
"inbound_did_audit",
"hunt_pilot_audit",
"whoami",
"did_block_overlap",
"partition_summary",
"dead_dn_finder",
"failed_fax_investigation",
}, f"unexpected prompt set: {names}"
# ---- new prompts: render-with-and-without-docs smoke tests -----------------
def test_phone_inventory_report_renders(fake_docs):
text = prompts.phone_inventory_report.render(fake_docs)
assert "Phone Inventory" in text
# Embedded SQL should query the `device` table with tkclass='Phone'
assert "tc.name = 'Phone'" in text
assert "all phones" in text
def test_phone_inventory_report_with_filter(fake_docs):
text = prompts.phone_inventory_report.render(fake_docs, "Lobby")
assert "%Lobby%" in text
assert "narrowed" in text
def test_phone_inventory_report_filter_escaped(fake_docs):
text = prompts.phone_inventory_report.render(fake_docs, "O'Hara")
assert "O''Hara" in text # SQL injection defense
def test_user_audit_default_focus(fake_docs):
text = prompts.user_audit.render(fake_docs)
assert "focus: `full`" in text
assert "applicationuser" in text
assert "enduser" in text
def test_user_audit_admin_focus(fake_docs):
text = prompts.user_audit.render(fake_docs, "admin")
assert "focus: `admin`" in text
def test_user_audit_unknown_focus_falls_back_to_full(fake_docs):
text = prompts.user_audit.render(fake_docs, "bogus_value")
assert "focus: `full`" in text
def test_inbound_did_audit_renders(fake_docs):
text = prompts.inbound_did_audit.render(fake_docs)
assert "Inbound DID Audit" in text
assert "XFORM-Inbound-DNIS" in text
assert "PSTN-Screen-PT" in text
def test_hunt_pilot_audit_uses_correct_column(fake_docs):
"""Hamilton bonus finding: huntpilotqueue joins via fknumplan_pilot,
NOT fknumplan. The prompt's embedded SQL must use the correct column
or every audit run will silently fail one category."""
text = prompts.hunt_pilot_audit.render(fake_docs)
assert "fknumplan_pilot" in text, (
"hunt_pilot_audit must use the verified column name `fknumplan_pilot`; "
"`fknumplan` was the silent-failure version we fixed in Hamilton review"
)
assert "tkpatternusage = 7" in text # Hunt Pilot type code
def test_all_new_prompts_render_without_docs():
"""Graceful degradation: each prompt produces usable output even when
the docs index isn't loaded."""
for name, fn, args in [
("phone_inventory_report", prompts.phone_inventory_report.render, ()),
("user_audit", prompts.user_audit.render, ()),
("inbound_did_audit", prompts.inbound_did_audit.render, ()),
("hunt_pilot_audit", prompts.hunt_pilot_audit.render, ()),
("whoami", prompts.whoami.render, ()),
]:
text = fn(None, *args)
assert "cisco-docs index is not loaded" in text, (
f"{name} failed graceful degradation"
)
# ---- whoami specifics ------------------------------------------------------
def test_whoami_uses_correct_table_names_for_cucm_15(fake_docs):
"""The query principle came in with `enduserauthgroupmap` and
`dirgrouprolemap`, but those tables don't exist on CUCM 15. The
prompt MUST embed the verified names: `enduserdirgroupmap` and
`functionroledirgroupmap`. If a future contributor reverts to the
older names, this test fires red."""
text = prompts.whoami.render(fake_docs, "TestUser")
assert "enduserdirgroupmap" in text
assert "functionroledirgroupmap" in text
# And the older deprecated names should NOT appear in the SQL we ship
# (they're fine in the prose section that explains why the names changed)
sql_section = text.split("## Schema knowledge")[0] + text.split("## Step")[1]
for old_name in ("enduserauthgroupmap", "dirgrouprolemap"):
assert old_name not in sql_section, (
f"deprecated table name {old_name} appeared in the SQL section"
)
def test_whoami_explicit_userid_in_query(fake_docs):
text = prompts.whoami.render(fake_docs, "SomeAccount")
assert "SomeAccount" in text
assert "explicit userid" in text
def test_whoami_default_uses_axl_user_env(fake_docs, monkeypatch):
monkeypatch.setenv("AXL_USER", "EnvAccount")
text = prompts.whoami.render(fake_docs)
assert "EnvAccount" in text
assert "AXL service account" in text
def test_whoami_no_userid_no_env(fake_docs, monkeypatch):
monkeypatch.delenv("AXL_USER", raising=False)
text = prompts.whoami.render(fake_docs)
assert "no userid supplied" in text
# Tells the LLM what to do — set the param or the env var
assert "set userid parameter" in text or "AXL_USER env var" in text
def test_whoami_userid_escaped_for_sql_safety(fake_docs):
text = prompts.whoami.render(fake_docs, "O'Brien")
assert "O''Brien" in text # single quote doubled per Informix convention