docs: align model-assisted examples with fail-closed configs

docs/examples.md

@@ -133,7 +133,7 @@ Treat these as manual/advisory runs, not default CI gates.
 
 ```bash
 pnpm exec eval-kit judge-coverage \
-  --config evals/eval-kit.config.json \
+  --config evals/eval-kit.model-judge.config.json \
   --case case-alpha \
   --candidate evals/cases/case-alpha/candidate.md \
   --model gpt-5.4 \

docs/guides/model-assisted-evals.md

@@ -55,26 +55,34 @@ the consumer has calibrated pointwise evidence and a specific comparison need.
 
 ## Commands
 
+Pointwise judging uses the standard manual model-judge config:
+
 ```bash
-pnpm exec eval-kit generate \
+pnpm exec eval-kit judge-coverage \
+  --config evals/eval-kit.model-judge.config.json \
   --case <case-id> \
+  --candidate <candidate.md> \
   --model <model> \
   --provider openai \
   --effort medium \
   --run-id <run-id>
+```
 
+Generation and pairwise are separate manual lanes. They require separate configs whose relevant
+method is explicitly enabled; do not run them through `evals/eval-kit.config.json` or
+`evals/eval-kit.model-judge.config.json`.
 
-pnpm exec eval-kit judge-coverage \
-  --config evals/eval-kit.model-judge.config.json \
+```bash
+pnpm exec eval-kit generate \
+  --config evals/eval-kit.generate.config.json \
   --case <case-id> \
-  --candidate <candidate.md> \
   --model <model> \
   --provider openai \
   --effort medium \
   --run-id <run-id>
 
-
 pnpm exec eval-kit judge-pairwise \
+  --config evals/eval-kit.pairwise.config.json \
   --case <case-id> \
   --candidate-a <a.md> \
   --candidate-b <b.md> \

docs/reference/cli.md

@@ -115,11 +115,13 @@ pnpm exec eval-kit generate \
   --provider <openai|openai:codex-app-server> \
   --effort <low|medium|high> \
   --run-id <id> \
-  [--config <path>]
+  --config evals/eval-kit.generate.config.json
 ```
 
 Requires Promptfoo and local Codex auth. This is manual/advisory evidence, not a default CI gate.
 Fails closed before those checks unless `methods.generate.enabled` is explicitly `true`.
+Use a separate explicit config for generation; do not enable generation in the deterministic default
+config.
 
 ## `judge-coverage`
 
@@ -133,7 +135,7 @@ pnpm exec eval-kit judge-coverage \
   --provider <openai|openai:codex-app-server> \
   --effort <low|medium|high> \
   [--run-id <id>] \
-  [--config <path>]
+  --config evals/eval-kit.model-judge.config.json
 ```
 
 Fails closed unless `methods.judge_coverage.enabled` is explicitly `true`.
@@ -157,10 +159,12 @@ pnpm exec eval-kit judge-pairwise \
   --effort <low|medium|high> \
   --seed <number> \
   --run-id <id> \
-  [--config <path>]
+  --config evals/eval-kit.pairwise.config.json
 ```
 
 Fails closed unless `methods.judge_pairwise.enabled` is explicitly `true`.
+Use a separate explicit config for pairwise after pointwise calibration; do not enable pairwise in
+the deterministic default config.
 
 ## `report`
 

docs/reference/results.md

@@ -39,7 +39,7 @@ Current schema:
   "run_type": "deterministic",
   "runner": {
     "id": "generic-eval-case",
-    "version": "0.1.2"
+    "version": "0.1.4"
   },
   "case_ids": ["case-example-v1"],
   "started_at": "2026-07-03T00:00:00.000Z",