Rename "MCP-unique" terminology to "org-scale" across codebase

Org suites are organizational use-case benchmarks (cross-repo discovery,
compliance, incident triage, migration, etc.) that both baseline and MCP
agents can solve — they are not MCP-dependent. This commit:

- Renames mcp_unique=true to org_scale=true in 241 task.toml files
- Updates 241 oracle_checks.py docstrings and 241 eval.sh comments
- Renames docs/MCP_UNIQUE_TASKS.md → docs/ORG_TASKS.md
- Renames docs/MCP_UNIQUE_CALIBRATION.md → docs/ORG_CALIBRATION.md
- Updates script variable names (MCP_UNIQUE_SUITES → ORG_SUITES, etc.)
- Adds 41 missing org tasks to selected_benchmark_tasks.json (T.7 PASS)
- All 20 suites remain Grade A with zero failures

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

Author: sjarmak

README.md

@@ -15,7 +15,7 @@ time_limit_sec = 900
 mcp_suite = "csb_org_compliance"
 use_case_id = 57
 repo_set_id = "grafana-observability"
-mcp_unique = true
+org_scale = true
 verification_modes = ["artifact"]
 
 [verification]

benchmarks/README.md

@@ -1,5 +1,5 @@
 #!/bin/bash
-# eval.sh — MCP-unique benchmark evaluator for CCX-compliance-057
+# eval.sh — org-scale benchmark evaluator for CCX-compliance-057
 # Exit-code-first (SWE-Factory pattern):
 #   exit 0 — agent produced useful output (composite score > 0)
 #   exit 1 — total failure (composite score == 0 or missing answer)

benchmarks/backups/csb_org_compliance_doe_trim/ccx-compliance-057/task.toml

@@ -1,5 +1,5 @@
 #!/usr/bin/env python3
-"""Deterministic oracle check library for MCP-unique benchmark evaluation.
+"""Deterministic oracle check library for org-scale benchmark evaluation.
 
 Provides reusable check functions that eval.sh scripts invoke to score agent
 answers against closed-world oracle definitions. Returns raw scores (no

benchmarks/backups/csb_org_compliance_doe_trim/ccx-compliance-057/tests/eval.sh

@@ -15,7 +15,7 @@ time_limit_sec = 900
 mcp_suite = "csb_org_compliance"
 use_case_id = 188
 repo_set_id = "envoy-service-mesh"
-mcp_unique = true
+org_scale = true
 verification_modes = ["artifact"]
 
 [verification]

benchmarks/backups/csb_org_compliance_doe_trim/ccx-compliance-057/tests/oracle_checks.py

@@ -1,5 +1,5 @@
 #!/bin/bash
-# eval.sh — MCP-unique benchmark evaluator for CCX-compliance-188
+# eval.sh — org-scale benchmark evaluator for CCX-compliance-188
 # Exit-code-first (SWE-Factory pattern):
 #   exit 0 — agent produced useful output (composite score > 0)
 #   exit 1 — total failure (composite score == 0 or missing answer)

benchmarks/backups/csb_org_compliance_doe_trim/ccx-compliance-188/task.toml

@@ -1,5 +1,5 @@
 #!/usr/bin/env python3
-"""Deterministic oracle check library for MCP-unique benchmark evaluation.
+"""Deterministic oracle check library for org-scale benchmark evaluation.
 
 Provides reusable check functions that eval.sh scripts invoke to score agent
 answers against closed-world oracle definitions. Returns raw scores (no

benchmarks/backups/csb_org_compliance_doe_trim/ccx-compliance-188/tests/eval.sh

@@ -15,7 +15,7 @@ time_limit_sec = 1200
 mcp_suite = "csb_org_compliance"
 use_case_id = 57
 repo_set_id = "grafana-observability"
-mcp_unique = true
+org_scale = true
 deepsearch_relevant = true
 
 [verification]

benchmarks/backups/csb_org_compliance_doe_trim/ccx-compliance-188/tests/oracle_checks.py

@@ -1,5 +1,5 @@
 #!/bin/bash
-# eval.sh — MCP-unique benchmark evaluator for CCX-compliance-057-ds
+# eval.sh — org-scale benchmark evaluator for CCX-compliance-057-ds
 # Exit-code-first (SWE-Factory pattern):
 #   exit 0 — agent produced useful output (composite score > 0)
 #   exit 1 — total failure (composite score == 0 or missing answer)