Restore tracker mapping linking, drop only create

The tracker-mapping worker had been reduced to catalog resolution only,
which removed not just the auto-creation of an org ThirdParty but also
the auto-linking of an existing one. Only the creation needed to go: it
raced the load-then-create check and produced duplicate vendors.

Restore the full org ThirdParty resolution (exact common-id link,
sibling direct-link, high-confidence heuristic, and the disambiguation
agent) and remove only the CreateFromCommon branch and its
categorisation gate. When nothing matches, the worker now leaves
third_party_id unset rather than creating a vendor; creation happens
exclusively through the explicit ImportFromCommon action. Drop the
now-dead CreateFromCommon helper and rename match.go to common_match.go.

Fix a latent test bug surfaced by actually running the DB-backed suite
(skipped in CI without Postgres): the heuristic-match candidate lacked
Level 1, so the level-filtered candidate loader excluded it and the old
fallback create masked the miss.

Signed-off-by: Émile Ré <emile@probo.com>

contrib/helm/charts/probo/CHANGELOG.md

@@ -4,10 +4,6 @@ All notable changes to the Probo Helm chart will be documented in this file.
 
 ## Unreleased
 
-### Removed
-
-- `thirdPartyDisambiguation` agent config slot and `trackerMappingWorker.disambiguationAgentTimeout`; the tracker-mapping worker no longer runs a third-party disambiguation agent
-
 ## [0.5.0] - 2026-06-09
 
 ### Added

contrib/helm/charts/probo/templates/deployment.yaml

@@ -281,6 +281,23 @@ spec:
             - name: THIRD_PARTY_VETTING_STALE_AFTER
               value: {{ .Values.probo.thirdPartyVettingWorker.staleAfter | quote }}
             {{- end }}
+            # Third-party Disambiguation Agent
+            {{- if .Values.probo.thirdPartyDisambiguation.provider }}
+            - name: AGENT_THIRD_PARTY_DISAMBIGUATION_PROVIDER
+              value: {{ .Values.probo.thirdPartyDisambiguation.provider | quote }}
+            {{- end }}
+            {{- if .Values.probo.thirdPartyDisambiguation.modelName }}
+            - name: AGENT_THIRD_PARTY_DISAMBIGUATION_MODEL_NAME
+              value: {{ .Values.probo.thirdPartyDisambiguation.modelName | quote }}
+            {{- end }}
+            {{- if .Values.probo.thirdPartyDisambiguation.temperature }}
+            - name: AGENT_THIRD_PARTY_DISAMBIGUATION_TEMPERATURE
+              value: {{ .Values.probo.thirdPartyDisambiguation.temperature | quote }}
+            {{- end }}
+            {{- if .Values.probo.thirdPartyDisambiguation.maxTokens }}
+            - name: AGENT_THIRD_PARTY_DISAMBIGUATION_MAX_TOKENS
+              value: {{ .Values.probo.thirdPartyDisambiguation.maxTokens | quote }}
+            {{- end }}
             # Tracker Mapping Agent
             {{- if .Values.probo.trackerMapping.provider }}
             - name: AGENT_TRACKER_MAPPING_PROVIDER
@@ -336,6 +353,10 @@ spec:
             - name: TRACKER_MAPPING_AGENT_MAX_TURNS
               value: {{ .Values.probo.trackerMappingWorker.agentMaxTurns | quote }}
             {{- end }}
+            {{- if .Values.probo.trackerMappingWorker.disambiguationAgentTimeout }}
+            - name: TRACKER_MAPPING_DISAMBIGUATION_AGENT_TIMEOUT
+              value: {{ .Values.probo.trackerMappingWorker.disambiguationAgentTimeout | quote }}
+            {{- end }}
             # Common Pattern Enrichment Worker
             {{- if .Values.probo.commonPatternEnrichmentWorker.interval }}
             - name: COMMON_PATTERN_ENRICHMENT_INTERVAL

contrib/helm/charts/probo/values-production.yaml.example

@@ -185,6 +185,15 @@ probo:
   #   maxConcurrency: 1
   #   staleAfter: 1500
 
+  # Third-party disambiguation agent (optional; runs inside the tracker
+  # mapping worker to pick the best matching org third party). Falls back
+  # to trackerMapping when its provider is unset.
+  # thirdPartyDisambiguation:
+  #   provider: "openai"
+  #   modelName: "gpt-4o"
+  #   temperature: "0.4"
+  #   maxTokens: "4096"
+
   # Tracker mapping agent (optional, auto-links tracker patterns to vendors)
   # trackerMapping:
   #   provider: "openai"
@@ -201,7 +210,7 @@ probo:
   #   maxTokens: "4096"
 
   # Tracker mapping worker tuning (optional; seconds for
-  # interval/staleAfter/agentTimeout).
+  # interval/staleAfter/agentTimeout/disambiguationAgentTimeout).
   # Keep concurrency modest to stay under OpenAI/Firecrawl limits and the DB pool.
   # trackerMappingWorker:
   #   interval: 10
@@ -209,6 +218,7 @@ probo:
   #   staleAfter: 600
   #   agentTimeout: 45
   #   agentMaxTurns: 10
+  #   disambiguationAgentTimeout: 45
 
   # Common-pattern enrichment worker tuning (optional; seconds for
   # interval/staleAfter/agentTimeout).

contrib/helm/charts/probo/values.yaml

@@ -286,6 +286,15 @@ probo:
     maxConcurrency: 1
     staleAfter: 1500
 
+  # Third-party disambiguation agent (optional; runs inside the tracker
+  # mapping worker to pick the best matching org third party). Falls back
+  # to trackerMapping when its provider is unset.
+  thirdPartyDisambiguation:
+    provider: ""
+    modelName: ""
+    temperature: ""
+    maxTokens: ""
+
   # Tracker mapping agent (optional, requires openai.apiKey or anthropic key)
   trackerMapping:
     provider: ""
@@ -302,15 +311,16 @@ probo:
     maxTokens: ""
 
   # Tracker mapping background worker tuning (optional). interval,
-  # staleAfter, and agentTimeout are in seconds. Keep concurrency modest
-  # to stay under OpenAI/Firecrawl rate limits and the database
-  # connection pool.
+  # staleAfter, agentTimeout, and disambiguationAgentTimeout are in
+  # seconds. Keep concurrency modest to stay under OpenAI/Firecrawl rate
+  # limits and the database connection pool.
   trackerMappingWorker:
     interval: 10
     maxConcurrency: 3
     staleAfter: 600
     agentTimeout: 45
     agentMaxTurns: 10
+    disambiguationAgentTimeout: 45
 
   # Common-pattern enrichment background worker tuning (optional).
   # interval, staleAfter, and agentTimeout are in seconds.