From 92a1a8ed3e84d87d26866a633c8abebb8c98d950 Mon Sep 17 00:00:00 2001
From: Seongho Bae <me@seonghobae.me>
Date: Fri, 26 Jun 2026 13:12:42 +0900
Subject: [PATCH 1/2] Prefer DeepSeek R1 for OpenCode reviews

---
 .github/workflows/opencode-review.yml | 113 ++++++++++++++++++--------
 opencode.jsonc                        |  45 +++++++++-
 scripts/ci/test_strix_quick_gate.sh   |  38 +++++----
 3 files changed, 142 insertions(+), 54 deletions(-)

diff --git a/.github/workflows/opencode-review.yml b/.github/workflows/opencode-review.yml
index ced41f6d4..9c13bb55c 100644
--- a/.github/workflows/opencode-review.yml
+++ b/.github/workflows/opencode-review.yml
@@ -881,7 +881,7 @@ jobs:
 
           jq -n --arg workspace "$OPENCODE_SOURCE_WORKDIR" '{
             "$schema": "https://opencode.ai/config.json",
-            "model": "github-models/openai/gpt-5",
+            "model": "github-models/deepseek/deepseek-r1-0528",
             "small_model": "github-models/deepseek/deepseek-v3-0324",
             "enabled_providers": ["github-models"],
             "lsp": true,
@@ -1002,6 +1002,24 @@ jobs:
                       "output": 100000
                     }
                   },
+                  "openai/gpt-5-chat": {
+                    "name": "OpenAI GPT-5 Chat",
+                    "tool_call": true,
+                    "reasoning": true,
+                    "limit": {
+                      "context": 200000,
+                      "output": 100000
+                    }
+                  },
+                  "openai/gpt-5-mini": {
+                    "name": "OpenAI GPT-5 Mini",
+                    "tool_call": true,
+                    "reasoning": true,
+                    "limit": {
+                      "context": 200000,
+                      "output": 100000
+                    }
+                  },
                   "deepseek/deepseek-r1-0528": {
                     "name": "DeepSeek R1 0528",
                     "tool_call": true,
@@ -1028,6 +1046,15 @@ jobs:
                       "output": 100000
                     }
                   },
+                  "openai/o3-mini": {
+                    "name": "OpenAI o3-mini",
+                    "tool_call": true,
+                    "reasoning": true,
+                    "limit": {
+                      "context": 200000,
+                      "output": 100000
+                    }
+                  },
                   "openai/o4-mini": {
                     "name": "OpenAI o4-mini",
                     "tool_call": true,
@@ -1036,6 +1063,22 @@ jobs:
                       "context": 200000,
                       "output": 100000
                     }
+                  },
+                  "mistral-ai/mistral-medium-2505": {
+                    "name": "Mistral Medium 3 25.05",
+                    "tool_call": true,
+                    "limit": {
+                      "context": 128000,
+                      "output": 4096
+                    }
+                  },
+                  "meta/llama-4-scout-17b-16e-instruct": {
+                    "name": "Llama 4 Scout 17B 16E Instruct",
+                    "tool_call": true,
+                    "limit": {
+                      "context": 1000000,
+                      "output": 4096
+                    }
                   }
                 }
               }
@@ -1044,21 +1087,21 @@ jobs:
 
           printf 'Prepared isolated OpenCode review workspace: %s\n' "$OPENCODE_REVIEW_WORKDIR"
 
-      - name: Run OpenCode PR Review (GPT-5)
+      - name: Run OpenCode PR Review (DeepSeek R1)
         id: opencode_review_primary
         if: needs.coverage-evidence.result == 'success'
         continue-on-error: true
-        timeout-minutes: 20
+        timeout-minutes: 15
         env:
           STRIX_GITHUB_MODELS_TOKEN: ${{ secrets.STRIX_GITHUB_MODELS_TOKEN }}
           GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
-          MODEL: github-models/openai/gpt-5
+          MODEL: github-models/deepseek/deepseek-r1-0528
           USE_GITHUB_TOKEN: "true"
           SHARE: "false"
           NPM_CONFIG_IGNORE_SCRIPTS: "true"
           NO_COLOR: "1"
-          OPENCODE_MODEL_ATTEMPTS: "1"
-          OPENCODE_RUN_TIMEOUT_SECONDS: "600"
+          OPENCODE_MODEL_ATTEMPTS: "3"
+          OPENCODE_RUN_TIMEOUT_SECONDS: "180"
           OPENCODE_EVIDENCE_FILE: ${{ runner.temp }}/opencode-review-evidence.md
           OPENCODE_OUTPUT_FILE: ${{ runner.temp }}/opencode-review-primary.md
           OPENCODE_REVIEW_WORKDIR: ${{ runner.temp }}/opencode-review-project
@@ -1171,7 +1214,7 @@ jobs:
           fi
           record_review_status "success"
 
-      - name: Run OpenCode PR Review fallback (DeepSeek R1)
+      - name: Run OpenCode PR Review fallback (DeepSeek V3)
         id: opencode_review_fallback
         if: >-
           always()
@@ -1182,7 +1225,7 @@ jobs:
         env:
           STRIX_GITHUB_MODELS_TOKEN: ${{ secrets.STRIX_GITHUB_MODELS_TOKEN }}
           GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
-          MODEL: github-models/deepseek/deepseek-r1-0528
+          MODEL: github-models/deepseek/deepseek-v3-0324
           USE_GITHUB_TOKEN: "true"
           SHARE: "false"
           NPM_CONFIG_IGNORE_SCRIPTS: "true"
@@ -1206,7 +1249,7 @@ jobs:
           }
           prompt_file="${RUNNER_TEMP}/opencode-review-prompt.md"
           cat >"$prompt_file" <<EOF
-          GPT-5 failed; review PR #${PR_NUMBER} in ${OPENCODE_SOURCE_WORKDIR} with DeepSeek R1-0528. The trusted workflow checkout is ${GITHUB_WORKSPACE}; inspect the pull request head source only from ${OPENCODE_SOURCE_WORKDIR}. Be general-purpose and meticulous: actively consult CodeGraph MCP for structural checks, DeepWiki for repo docs, Context7 for current library/API docs, and web_search for bounded external lookups such as action/tool release facts, industry standards, international standards, official platform specifications, and comparable issue or PR precedents when applicable. Do not rely on model memory for user-claimed concepts, standards, runtime support, or domain terminology when a search source is available. If a configured MCP source is unavailable or not applicable, say so briefly in the review summary. Inspect changed files and focused hunks directly when MCP evidence is insufficient. Do not claim repository docs, images, or reference assets are unavailable, missing, or absent unless the changed docs repository tree evidence proves it. If an external MCP source is unavailable, state that as a source limitation, not as a repository fact.
+          DeepSeek R1-0528 failed; review PR #${PR_NUMBER} in ${OPENCODE_SOURCE_WORKDIR} with DeepSeek V3-0324. The trusted workflow checkout is ${GITHUB_WORKSPACE}; inspect the pull request head source only from ${OPENCODE_SOURCE_WORKDIR}. Be general-purpose and meticulous: actively consult CodeGraph MCP for structural checks, DeepWiki for repo docs, Context7 for current library/API docs, and web_search for bounded external lookups such as action/tool release facts, industry standards, international standards, official platform specifications, and comparable issue or PR precedents when applicable. Do not rely on model memory for user-claimed concepts, standards, runtime support, or domain terminology when a search source is available. If a configured MCP source is unavailable or not applicable, say so briefly in the review summary. Inspect changed files and focused hunks directly when MCP evidence is insufficient. Do not claim repository docs, images, or reference assets are unavailable, missing, or absent unless the changed docs repository tree evidence proves it. If an external MCP source is unavailable, state that as a source limitation, not as a repository fact.
           Structural exploration is mandatory for every PR, including dependency-only, lockfile-only, workflow-only, docs-only, and no-source-code changes; inspect the relevant manifest, lockfile, workflow, config, docs, dependency edges, generated side effects, code-to-documentation consistency, documentation-to-code consistency, and test-command contracts. Docs-only changes still require CodeGraph, DeepWiki, Context7, or web_search evidence when they make claims about behavior, APIs, setup, workflows, dependencies, standards, or product/domain concepts. If changed documentation contradicts current code, generated behavior, official docs, repository docs, or reachable standards evidence, request changes with a source-backed fix direction: either fix the documentation claim or update the code/contract that makes the claim false. Never state that structural exploration, structural analysis, or structural review is not required or unnecessary. If structural exploration was not possible or changed files could not be inspected after reading bounded-review-evidence.md and the changed files, do not approve. If evidence is truncated, inspect focused hunks and changed files directly before deciding. Do not request changes solely because the prompt did not inline the full evidence.
           Use CodeGraph for blast-radius, call graph, and test-coverage questions before broad local reads. Prefer deletion, stdlib/native platform features, and already-installed dependencies before proposing new code or packages, but do not simplify away trust-boundary validation, data-loss handling, security, accessibility, or required tests. Follow the Review language evidence section: write human-readable review prose in Korean when the PR title or body is primarily Korean, and in English when it is primarily English. Keep file paths, code identifiers, commands, logs, quoted source, error text, numbers, and protocol literals unchanged. For Korean prose, preserve facts, identifiers, numbers, and quotes while removing only formulaic filler or translationese.
           Cover security/privacy boundaries, tenant isolation, workflow contracts, developer experience, user-facing behavior, tests, cross-file compatibility, repository conventions, and regression risk. Compare repository-local patterns before judging DX or UX: preserve helpful automation, review, setup, documentation, and product-flow patterns from sibling repositories when they reduce cognitive load or user friction, and flag patterns that only add noise, false failures, misleading status, repeated waiting, or URL-only diagnostics. For schema, migration, database, API, workflow, security, or compliance changes, compare against nearby implementation, code conventions, reserved words, naming rules, applicable standards, git history, and deployment evidence before approving. For breaking changes, especially when bounded evidence shows production deployment records, comment on backward-compatibility impact, migration or bridge-module needs, rollout/rollback path, and lower-version compatibility.
@@ -1254,7 +1297,7 @@ jobs:
             fi
           done
           if [ "$opencode_run_status" -ne 0 ]; then
-            echo "OpenCode DeepSeek R1 review attempt did not complete; next fallback review will run."
+            echo "OpenCode DeepSeek V3 review attempt did not complete; next fallback review will run."
             record_review_status "failed"
             exit 0
           fi
@@ -1301,7 +1344,7 @@ jobs:
           fi
           record_review_status "success"
 
-      - name: Run OpenCode PR Review fallback (DeepSeek V3)
+      - name: Run OpenCode PR Review fallback (GPT-5)
         id: opencode_review_second_fallback
         if: >-
           always()
@@ -1309,16 +1352,16 @@ jobs:
           && steps.opencode_review_primary.outputs.review_status != 'success'
           && steps.opencode_review_fallback.outputs.review_status != 'success'
         continue-on-error: true
-        timeout-minutes: 15
+        timeout-minutes: 8
         env:
           STRIX_GITHUB_MODELS_TOKEN: ${{ secrets.STRIX_GITHUB_MODELS_TOKEN }}
           GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
-          MODEL: github-models/deepseek/deepseek-v3-0324
+          MODEL: github-models/openai/gpt-5
           USE_GITHUB_TOKEN: "true"
           SHARE: "false"
           NPM_CONFIG_IGNORE_SCRIPTS: "true"
           NO_COLOR: "1"
-          OPENCODE_MODEL_ATTEMPTS: "3"
+          OPENCODE_MODEL_ATTEMPTS: "1"
           OPENCODE_RUN_TIMEOUT_SECONDS: "180"
           OPENCODE_EVIDENCE_FILE: ${{ runner.temp }}/opencode-review-evidence.md
           OPENCODE_OUTPUT_FILE: ${{ runner.temp }}/opencode-review-second-fallback.md
@@ -1337,7 +1380,7 @@ jobs:
           }
           prompt_file="${RUNNER_TEMP}/opencode-review-prompt.md"
           cat >"$prompt_file" <<EOF
-          GPT-5 and DeepSeek R1-0528 failed; review PR #${PR_NUMBER} in ${OPENCODE_SOURCE_WORKDIR} with DeepSeek V3-0324. The trusted workflow checkout is ${GITHUB_WORKSPACE}; inspect the pull request head source only from ${OPENCODE_SOURCE_WORKDIR}. Be general-purpose and meticulous: actively consult CodeGraph MCP for structural checks, DeepWiki for repo docs, Context7 for current library/API docs, and web_search for bounded external lookups such as action/tool release facts, industry standards, international standards, official platform specifications, and comparable issue or PR precedents when applicable. Do not rely on model memory for user-claimed concepts, standards, runtime support, or domain terminology when a search source is available. If a configured MCP source is unavailable or not applicable, say so briefly in the review summary. Inspect changed files and focused hunks directly when MCP evidence is insufficient. Do not claim repository docs, images, or reference assets are unavailable, missing, or absent unless the changed docs repository tree evidence proves it. If an external MCP source is unavailable, state that as a source limitation, not as a repository fact.
+          DeepSeek R1-0528 and DeepSeek V3-0324 failed; review PR #${PR_NUMBER} in ${OPENCODE_SOURCE_WORKDIR} with GPT-5. The trusted workflow checkout is ${GITHUB_WORKSPACE}; inspect the pull request head source only from ${OPENCODE_SOURCE_WORKDIR}. Be general-purpose and meticulous: actively consult CodeGraph MCP for structural checks, DeepWiki for repo docs, Context7 for current library/API docs, and web_search for bounded external lookups such as action/tool release facts, industry standards, international standards, official platform specifications, and comparable issue or PR precedents when applicable. Do not rely on model memory for user-claimed concepts, standards, runtime support, or domain terminology when a search source is available. If a configured MCP source is unavailable or not applicable, say so briefly in the review summary. Inspect changed files and focused hunks directly when MCP evidence is insufficient. Do not claim repository docs, images, or reference assets are unavailable, missing, or absent unless the changed docs repository tree evidence proves it. If an external MCP source is unavailable, state that as a source limitation, not as a repository fact.
           Structural exploration is mandatory for every PR, including dependency-only, lockfile-only, workflow-only, docs-only, and no-source-code changes; inspect the relevant manifest, lockfile, workflow, config, docs, dependency edges, generated side effects, code-to-documentation consistency, documentation-to-code consistency, and test-command contracts. Docs-only changes still require CodeGraph, DeepWiki, Context7, or web_search evidence when they make claims about behavior, APIs, setup, workflows, dependencies, standards, or product/domain concepts. If changed documentation contradicts current code, generated behavior, official docs, repository docs, or reachable standards evidence, request changes with a source-backed fix direction: either fix the documentation claim or update the code/contract that makes the claim false. Never state that structural exploration, structural analysis, or structural review is not required or unnecessary. If structural exploration was not possible or changed files could not be inspected after reading bounded-review-evidence.md and the changed files, do not approve. If evidence is truncated, inspect focused hunks and changed files directly before deciding. Do not request changes solely because the prompt did not inline the full evidence.
           Use CodeGraph for blast-radius, call graph, and test-coverage questions before broad local reads. Prefer deletion, stdlib/native platform features, and already-installed dependencies before proposing new code or packages, but do not simplify away trust-boundary validation, data-loss handling, security, accessibility, or required tests. Follow the Review language evidence section: write human-readable review prose in Korean when the PR title or body is primarily Korean, and in English when it is primarily English. Keep file paths, code identifiers, commands, logs, quoted source, error text, numbers, and protocol literals unchanged. For Korean prose, preserve facts, identifiers, numbers, and quotes while removing only formulaic filler or translationese.
           Cover security/privacy boundaries, tenant isolation, workflow contracts, developer experience, user-facing behavior, tests, cross-file compatibility, repository conventions, and regression risk. Compare repository-local patterns before judging DX or UX: preserve helpful automation, review, setup, documentation, and product-flow patterns from sibling repositories when they reduce cognitive load or user friction, and flag patterns that only add noise, false failures, misleading status, repeated waiting, or URL-only diagnostics. For schema, migration, database, API, workflow, security, or compliance changes, compare against nearby implementation, code conventions, reserved words, naming rules, applicable standards, git history, and deployment evidence before approving. For breaking changes, especially when bounded evidence shows production deployment records, comment on backward-compatibility impact, migration or bridge-module needs, rollout/rollback path, and lower-version compatibility.
@@ -1385,7 +1428,7 @@ jobs:
             fi
           done
           if [ "$opencode_run_status" -ne 0 ]; then
-            echo "OpenCode DeepSeek V3 review attempt did not complete."
+            echo "OpenCode GPT-5 review attempt did not complete."
             record_review_status "failed"
             exit 0
           fi
@@ -1432,8 +1475,8 @@ jobs:
           fi
           record_review_status "success"
 
-      - name: Run OpenCode PR Review fallback (OpenAI o-series)
-        id: opencode_review_o_series_fallback
+      - name: Run OpenCode PR Review fallback (catalog model pool)
+        id: opencode_review_catalog_fallback
         if: >-
           always()
           && needs.coverage-evidence.result == 'success'
@@ -1449,11 +1492,11 @@ jobs:
           SHARE: "false"
           NPM_CONFIG_IGNORE_SCRIPTS: "true"
           NO_COLOR: "1"
-          OPENCODE_MODEL_CANDIDATES: "github-models/openai/o3 github-models/openai/o4-mini"
+          OPENCODE_MODEL_CANDIDATES: "github-models/openai/gpt-5-chat github-models/openai/gpt-5-mini github-models/openai/o3 github-models/openai/o3-mini github-models/openai/o4-mini github-models/mistral-ai/mistral-medium-2505 github-models/meta/llama-4-scout-17b-16e-instruct"
           OPENCODE_MODEL_ATTEMPTS: "2"
           OPENCODE_RUN_TIMEOUT_SECONDS: "180"
           OPENCODE_EVIDENCE_FILE: ${{ runner.temp }}/opencode-review-evidence.md
-          OPENCODE_OUTPUT_FILE: ${{ runner.temp }}/opencode-review-o-series-fallback.md
+          OPENCODE_OUTPUT_FILE: ${{ runner.temp }}/opencode-review-catalog-fallback.md
           OPENCODE_REVIEW_WORKDIR: ${{ runner.temp }}/opencode-review-project
           OPENCODE_SOURCE_WORKDIR: ${{ runner.temp }}/opencode-pr-head
           PR_NUMBER: ${{ github.event.pull_request.number || github.event.inputs.pr_number }}
@@ -1494,7 +1537,7 @@ jobs:
             opencode_export_file="${candidate_output_file}.session.json"
             prompt_file="${RUNNER_TEMP}/opencode-review-${model_candidate//\//-}-prompt.md"
             cat >"$prompt_file" <<EOF
-          GPT-5, DeepSeek R1, and DeepSeek V3 did not produce a usable review. Review PR #${PR_NUMBER} in ${OPENCODE_SOURCE_WORKDIR} with ${model_candidate}. The trusted workflow checkout is ${GITHUB_WORKSPACE}; inspect the pull request head source only from ${OPENCODE_SOURCE_WORKDIR}.
+          DeepSeek R1-0528, DeepSeek V3-0324, and GPT-5 did not produce a usable review. Review PR #${PR_NUMBER} in ${OPENCODE_SOURCE_WORKDIR} with ${model_candidate}. The trusted workflow checkout is ${GITHUB_WORKSPACE}; inspect the pull request head source only from ${OPENCODE_SOURCE_WORKDIR}.
           CodeGraph MCP is mandatory for structural checks. Also use DeepWiki for repo docs, Context7 for current library/API docs, and web_search for bounded external lookups such as user-claimed concepts, industry standards, international standards, official platform specifications, and comparable issue or PR precedents when applicable. Do not rely on model memory for concepts, standards, runtime support, or domain terminology when a search source is available.
           Read ./bounded-review-evidence.md first, follow its Review language evidence for the final review language, then inspect changed files and focused hunks under the PR head worktree. Cover security/privacy boundaries, tenant isolation, workflow contracts, developer experience, user-facing behavior, tests, cross-file compatibility, repository conventions, deployment evidence, git history, breaking-change/backcompat impact, and regression risk. Compare repository-local patterns before judging DX or UX: preserve helpful automation, review, setup, documentation, and product-flow patterns from sibling repositories when they reduce cognitive load or user friction, and flag patterns that only add noise, false failures, misleading status, repeated waiting, or URL-only diagnostics. For schema, migration, database, API, workflow, security, or compliance changes, compare against nearby implementation, code conventions, reserved words, naming rules, applicable standards, and production deployment evidence before approving.
           Before APPROVE, the summary must name at least one exact changed file path and include a Verification posture section with these exact labels: Linter/static:, TDD/regression:, Coverage:, Docstring coverage:, DAG:, PoC/execution:, DDD/domain:, CDD/context:, Similar issues:, Claim/concept check:, Standards search:, Compatibility/convention:, Breaking-change/backcompat:, Performance:, Developer experience:, User experience:, Security/privacy:. The CDD/context label must explicitly mention CodeGraph or structural MCP evidence. Coverage and Docstring coverage labels must cite Coverage execution evidence proving 100%; missing, partial, skipped, unavailable, or not-applicable measurement is a blocker, not an approval condition. DAG: must name the rendered Change Flow DAG or an equivalent Mermaid DAG that maps changed files to affected execution path, main risk, and verification path. Developer experience: must state whether the change helps or obstructs maintainers, reviewers, CI operators, and future contributors, citing concrete repository evidence. User experience: must state whether product, documentation, review-comment, or status-check readers get clearer or worse outcomes, citing concrete evidence. PoC/execution: must cite the scratch proof, repro, focused test, lint, security, performance, or UI verification command that was actually run and its result; if no meaningful PoC can be run, state the exact repository limitation and request changes when the claim cannot otherwise be proven. If a surface is not applicable or unavailable, say why using that label.
@@ -1515,7 +1558,7 @@ jobs:
                 --agent ci-review-fallback \
                 --model "$model_candidate" \
                 --format json \
-                --title "PR #${PR_NUMBER} OpenCode bounded o-series review ${model_candidate} attempt ${opencode_attempt}/${opencode_attempts}" >"$opencode_json_file"
+                --title "PR #${PR_NUMBER} OpenCode bounded catalog fallback review ${model_candidate} attempt ${opencode_attempt}/${opencode_attempts}" >"$opencode_json_file"
               opencode_run_status=$?
               set -e
               if [ "$opencode_run_status" -ne 0 ]; then
@@ -1628,7 +1671,7 @@ jobs:
           && (steps.opencode_review_primary.outputs.review_status == 'success'
               || steps.opencode_review_fallback.outputs.review_status == 'success'
               || steps.opencode_review_second_fallback.outputs.review_status == 'success'
-              || steps.opencode_review_o_series_fallback.outputs.review_status == 'success')
+              || steps.opencode_review_catalog_fallback.outputs.review_status == 'success')
         env:
           GH_TOKEN: ${{ steps.opencode_app_token.outputs.token || secrets.OPENCODE_APPROVE_TOKEN || secrets.GITHUB_TOKEN }}
           GH_REPOSITORY: ${{ github.repository }}
@@ -1639,11 +1682,11 @@ jobs:
           OPENCODE_PRIMARY_OUTCOME: ${{ steps.opencode_review_primary.outputs.review_status }}
           OPENCODE_FALLBACK_OUTCOME: ${{ steps.opencode_review_fallback.outputs.review_status }}
           OPENCODE_SECOND_FALLBACK_OUTCOME: ${{ steps.opencode_review_second_fallback.outputs.review_status }}
-          OPENCODE_O_SERIES_FALLBACK_OUTCOME: ${{ steps.opencode_review_o_series_fallback.outputs.review_status }}
+          OPENCODE_CATALOG_FALLBACK_OUTCOME: ${{ steps.opencode_review_catalog_fallback.outputs.review_status }}
           OPENCODE_PRIMARY_OUTPUT_FILE: ${{ runner.temp }}/opencode-review-primary.md
           OPENCODE_FALLBACK_OUTPUT_FILE: ${{ runner.temp }}/opencode-review-fallback.md
           OPENCODE_SECOND_FALLBACK_OUTPUT_FILE: ${{ runner.temp }}/opencode-review-second-fallback.md
-          OPENCODE_O_SERIES_FALLBACK_OUTPUT_FILE: ${{ runner.temp }}/opencode-review-o-series-fallback.md
+          OPENCODE_CATALOG_FALLBACK_OUTPUT_FILE: ${{ runner.temp }}/opencode-review-catalog-fallback.md
           # The publish gate re-runs source-backed validation against PR-head data.
           OPENCODE_SOURCE_WORKDIR: ${{ runner.temp }}/opencode-pr-head
           PR_BASE_SHA: ${{ github.event.pull_request.base.sha || github.event.inputs.pr_base_sha }}
@@ -1658,7 +1701,7 @@ jobs:
           elif [ "$OPENCODE_SECOND_FALLBACK_OUTCOME" = "success" ]; then
             review_output_file="$OPENCODE_SECOND_FALLBACK_OUTPUT_FILE"
           else
-            review_output_file="$OPENCODE_O_SERIES_FALLBACK_OUTPUT_FILE"
+            review_output_file="$OPENCODE_CATALOG_FALLBACK_OUTPUT_FILE"
           fi
 
           clean_output="$(mktemp)"
@@ -1901,7 +1944,7 @@ jobs:
           COVERAGE_EVIDENCE_SUMMARY: ${{ needs.coverage-evidence.outputs.coverage_summary || 'Coverage evidence job did not run or did not publish coverage evidence.' }}
           OPENCODE_REVIEW_WORKDIR: ${{ runner.temp }}/opencode-review-project
           OPENCODE_SOURCE_WORKDIR: ${{ runner.temp }}/opencode-pr-head
-          MODEL: github-models/openai/gpt-5
+          MODEL: github-models/deepseek/deepseek-r1-0528
           USE_GITHUB_TOKEN: "true"
           NPM_CONFIG_IGNORE_SCRIPTS: "true"
           NO_COLOR: "1"
@@ -1912,11 +1955,11 @@ jobs:
           OPENCODE_PRIMARY_OUTCOME: ${{ steps.opencode_review_primary.outputs.review_status }}
           OPENCODE_FALLBACK_OUTCOME: ${{ steps.opencode_review_fallback.outputs.review_status }}
           OPENCODE_SECOND_FALLBACK_OUTCOME: ${{ steps.opencode_review_second_fallback.outputs.review_status }}
-          OPENCODE_O_SERIES_FALLBACK_OUTCOME: ${{ steps.opencode_review_o_series_fallback.outputs.review_status }}
+          OPENCODE_CATALOG_FALLBACK_OUTCOME: ${{ steps.opencode_review_catalog_fallback.outputs.review_status }}
           OPENCODE_PRIMARY_OUTPUT_FILE: ${{ runner.temp }}/opencode-review-primary.md
           OPENCODE_FALLBACK_OUTPUT_FILE: ${{ runner.temp }}/opencode-review-fallback.md
           OPENCODE_SECOND_FALLBACK_OUTPUT_FILE: ${{ runner.temp }}/opencode-review-second-fallback.md
-          OPENCODE_O_SERIES_FALLBACK_OUTPUT_FILE: ${{ runner.temp }}/opencode-review-o-series-fallback.md
+          OPENCODE_CATALOG_FALLBACK_OUTPUT_FILE: ${{ runner.temp }}/opencode-review-catalog-fallback.md
           PR_BASE_SHA: ${{ github.event.pull_request.base.sha || github.event.inputs.pr_base_sha }}
           PR_HEAD_SHA: ${{ github.event.pull_request.head.sha || github.event.inputs.pr_head_sha }}
           APPROVAL_CHECK_WAIT_ATTEMPTS: "81"
@@ -2670,8 +2713,8 @@ jobs:
               ".github/workflows/strix.yml" \
               "scripts/ci/test_strix_quick_gate.sh"
             emit_known_missing_string_finding \
-              "MODEL: github-models/openai/gpt-5" \
-              "OpenCode review must try GitHub Models GPT-5 first" \
+              "MODEL: github-models/deepseek/deepseek-r1-0528" \
+              "OpenCode review must start with DeepSeek R1" \
               ".github/workflows/opencode-review.yml" \
               "scripts/ci/test_strix_quick_gate.sh"
 
@@ -3527,7 +3570,7 @@ jobs:
             opencode_review_outcome="${OPENCODE_SECOND_FALLBACK_OUTCOME:-unknown}"
           fi
           if [ "$opencode_review_outcome" != "success" ]; then
-            opencode_review_outcome="${OPENCODE_O_SERIES_FALLBACK_OUTCOME:-unknown}"
+            opencode_review_outcome="${OPENCODE_CATALOG_FALLBACK_OUTCOME:-unknown}"
           fi
 
           if [ "$opencode_review_outcome" != "success" ]; then
@@ -3577,7 +3620,7 @@ jobs:
                 "all configured OpenCode model attempts failed to produce a usable current-head control block." \
                 "" \
                 "- Result: OPENCODE_REVIEW_UNAVAILABLE" \
-                "- Reason: OpenCode action outcomes were primary=${OPENCODE_PRIMARY_OUTCOME:-unknown}, fallback=${OPENCODE_FALLBACK_OUTCOME:-unknown}, second_fallback=${OPENCODE_SECOND_FALLBACK_OUTCOME:-unknown}, o_series_fallback=${OPENCODE_O_SERIES_FALLBACK_OUTCOME:-unknown}." \
+                "- Reason: OpenCode action outcomes were primary=${OPENCODE_PRIMARY_OUTCOME:-unknown}, fallback=${OPENCODE_FALLBACK_OUTCOME:-unknown}, second_fallback=${OPENCODE_SECOND_FALLBACK_OUTCOME:-unknown}, catalog_fallback=${OPENCODE_CATALOG_FALLBACK_OUTCOME:-unknown}." \
                 "- Required next evidence: rerun OpenCode with a model/tooling attempt that emits a valid source-backed control block for this head." \
                 "- Head SHA: \`${HEAD_SHA}\`" \
                 "- Workflow run: ${RUN_ID}" \
@@ -3597,8 +3640,8 @@ jobs:
             selected_review_output_file="${OPENCODE_FALLBACK_OUTPUT_FILE}"
           elif [ "${OPENCODE_SECOND_FALLBACK_OUTCOME:-}" = "success" ]; then
             selected_review_output_file="${OPENCODE_SECOND_FALLBACK_OUTPUT_FILE}"
-          elif [ "${OPENCODE_O_SERIES_FALLBACK_OUTCOME:-}" = "success" ]; then
-            selected_review_output_file="${OPENCODE_O_SERIES_FALLBACK_OUTPUT_FILE}"
+          elif [ "${OPENCODE_CATALOG_FALLBACK_OUTCOME:-}" = "success" ]; then
+            selected_review_output_file="${OPENCODE_CATALOG_FALLBACK_OUTPUT_FILE}"
           fi
 
           load_selected_review_output() {
diff --git a/opencode.jsonc b/opencode.jsonc
index 06e68e7c0..8adeee0e1 100644
--- a/opencode.jsonc
+++ b/opencode.jsonc
@@ -1,6 +1,6 @@
 {
   "$schema": "https://opencode.ai/config.json",
-  "model": "github-models/openai/gpt-5",
+  "model": "github-models/deepseek/deepseek-r1-0528",
   "small_model": "github-models/deepseek/deepseek-v3-0324",
   "enabled_providers": ["github-models"],
   "lsp": true,
@@ -107,6 +107,24 @@
             "output": 100000
           }
         },
+        "openai/gpt-5-chat": {
+          "name": "OpenAI GPT-5 Chat",
+          "tool_call": true,
+          "reasoning": true,
+          "limit": {
+            "context": 200000,
+            "output": 100000
+          }
+        },
+        "openai/gpt-5-mini": {
+          "name": "OpenAI GPT-5 Mini",
+          "tool_call": true,
+          "reasoning": true,
+          "limit": {
+            "context": 200000,
+            "output": 100000
+          }
+        },
         "deepseek/deepseek-r1-0528": {
           "name": "DeepSeek R1 0528",
           "tool_call": true,
@@ -133,6 +151,15 @@
             "output": 100000
           }
         },
+        "openai/o3-mini": {
+          "name": "OpenAI o3-mini",
+          "tool_call": true,
+          "reasoning": true,
+          "limit": {
+            "context": 200000,
+            "output": 100000
+          }
+        },
         "openai/o4-mini": {
           "name": "OpenAI o4-mini",
           "tool_call": true,
@@ -141,6 +168,22 @@
             "context": 200000,
             "output": 100000
           }
+        },
+        "mistral-ai/mistral-medium-2505": {
+          "name": "Mistral Medium 3 25.05",
+          "tool_call": true,
+          "limit": {
+            "context": 128000,
+            "output": 4096
+          }
+        },
+        "meta/llama-4-scout-17b-16e-instruct": {
+          "name": "Llama 4 Scout 17B 16E Instruct",
+          "tool_call": true,
+          "limit": {
+            "context": 1000000,
+            "output": 4096
+          }
         }
       }
     }
diff --git a/scripts/ci/test_strix_quick_gate.sh b/scripts/ci/test_strix_quick_gate.sh
index 37c57c8db..e67f07c16 100755
--- a/scripts/ci/test_strix_quick_gate.sh
+++ b/scripts/ci/test_strix_quick_gate.sh
@@ -447,14 +447,14 @@ assert_opencode_review_uses_codegraph_and_gpt5_fallback() {
 	assert_file_contains "$workflow_file" "Do not spend the session listing every changed path before reviewing" "opencode review prompt prevents fallback sessions from exhausting steps on file listing"
 	assert_file_contains "$workflow_file" "Always return a final control block instead of a progress summary" "opencode review prompt requires a gate conclusion instead of a progress summary"
 	assert_file_contains "$workflow_file" 'timeout --kill-after=30s "${OPENCODE_RUN_TIMEOUT_SECONDS:-180}s" opencode run' "opencode review primary model has a kill-after bounded timeout so fallback review can publish promptly"
-	assert_file_contains "$workflow_file" 'OPENCODE_RUN_TIMEOUT_SECONDS: "600"' "opencode primary review has enough bounded time for tool-backed current-head review"
+	assert_file_contains "$workflow_file" 'OPENCODE_RUN_TIMEOUT_SECONDS: "180"' "opencode primary review is bounded tightly enough to reach fallback models promptly"
 	assert_file_contains "$workflow_file" "&& needs.coverage-evidence.result == 'success'" "opencode model fallbacks only run after coverage evidence passed"
-	assert_file_contains "$workflow_file" "&& steps.opencode_review_primary.outputs.review_status != 'success'" "opencode DeepSeek R1 fallback still runs after a primary model timeout or step failure when coverage evidence passed"
+	assert_file_contains "$workflow_file" "&& steps.opencode_review_primary.outputs.review_status != 'success'" "opencode DeepSeek V3 fallback still runs after a primary model timeout or step failure when coverage evidence passed"
 	assert_file_contains "$workflow_file" "always()" "opencode fallback chain uses always() so failed model steps cannot skip every fallback"
-	assert_file_contains "$workflow_file" 'OPENCODE_MODEL_ATTEMPTS: "1"' "opencode primary review uses one longer attempt before exhausting the primary model"
-	assert_file_contains "$workflow_file" "Run OpenCode PR Review fallback (OpenAI o-series)" "opencode review includes extra reasoning-model fallback"
+	assert_file_contains "$workflow_file" 'OPENCODE_MODEL_ATTEMPTS: "1"' "opencode GPT-5 fallback uses one bounded attempt before trying the catalog pool"
+	assert_file_contains "$workflow_file" "Run OpenCode PR Review fallback (catalog model pool)" "opencode review includes a broad catalog fallback pool"
 	assert_file_contains "$workflow_file" "continue-on-error: true" "opencode model step timeouts do not prevent fallback review publication"
-	assert_file_contains "$workflow_file" "github-models/openai/o3 github-models/openai/o4-mini" "opencode review tries o-series reasoning models after GPT-5 and DeepSeek fallbacks"
+	assert_file_contains "$workflow_file" "github-models/openai/gpt-5-chat github-models/openai/gpt-5-mini github-models/openai/o3 github-models/openai/o3-mini github-models/openai/o4-mini github-models/mistral-ai/mistral-medium-2505 github-models/meta/llama-4-scout-17b-16e-instruct" "opencode review tries catalog-available tool-calling fallbacks after DeepSeek and GPT-5 paths"
 	assert_file_contains "$workflow_file" "The publish gate re-runs source-backed validation against PR-head data" "opencode review publish gate validates model output against the PR-head worktree"
 	assert_file_contains "$workflow_file" '"openai/o3"' "opencode config declares OpenAI o3 fallback"
 	assert_file_contains "$workflow_file" '"openai/o4-mini"' "opencode config declares OpenAI o4-mini fallback"
@@ -524,9 +524,9 @@ assert_opencode_review_uses_codegraph_and_gpt5_fallback() {
 	assert_file_contains "$workflow_file" "all configured OpenCode model attempts failed to produce a usable current-head control block" "opencode model-output failures fail the check without publishing a review"
 	assert_file_contains "$workflow_file" "Leaving the PR review unchanged because this is review tooling instability, not a source-code finding." "opencode model-failure path avoids PR review noise"
 	assert_file_contains "$workflow_file" 'OPENCODE_MODEL_ATTEMPTS: "3"' "opencode primary and deepseek review paths retry model execution"
-	assert_file_contains "$workflow_file" 'OPENCODE_MODEL_ATTEMPTS: "2"' "opencode o-series fallback retries each reasoning model"
-	assert_file_contains "$workflow_file" "OpenCode %s fallback attempt %s/%s failed" "opencode o-series fallback records per-model retry failures"
-	assert_file_contains "$workflow_file" "github-models/openai/o3 github-models/openai/o4-mini" "opencode review includes additional OpenAI reasoning model fallbacks"
+	assert_file_contains "$workflow_file" 'OPENCODE_MODEL_ATTEMPTS: "2"' "opencode catalog fallback retries each model"
+	assert_file_contains "$workflow_file" "OpenCode %s fallback attempt %s/%s failed" "opencode catalog fallback records per-model retry failures"
+	assert_file_contains "$workflow_file" "github-models/openai/o3 github-models/openai/o3-mini github-models/openai/o4-mini" "opencode review includes additional OpenAI reasoning model fallbacks"
 	assert_file_contains "$workflow_file" "coverage-evidence:" "opencode workflow measures coverage before review"
 	assert_file_contains "$workflow_file" "github.event_name == 'workflow_dispatch'" "manual current-head OpenCode reviews measure coverage instead of approving skipped coverage evidence"
 	assert_file_contains "$workflow_file" "if: github.event_name == 'workflow_dispatch'" "pull_request_target must not execute PR-head coverage scripts"
@@ -617,9 +617,10 @@ assert_opencode_review_uses_codegraph_and_gpt5_fallback() {
 	assert_file_not_contains "$workflow_file" 'repos/${{ github.repository }}' "opencode review workflow must pass repository expressions through env before shell use"
 	assert_file_contains "$workflow_file" "GH_REPOSITORY:" "opencode review workflow exports repository context through env"
 	assert_file_contains "$workflow_file" 'repos/${GH_REPOSITORY}' "opencode review workflow uses env-backed repository context in shell commands"
-	assert_file_contains "$workflow_file" "MODEL: github-models/openai/gpt-5" "opencode review tries GitHub Models GPT-5 first"
-	assert_file_contains "$workflow_file" "MODEL: github-models/deepseek/deepseek-r1-0528" "opencode review falls back to a reachable DeepSeek R1 reasoning model"
-	assert_file_contains "$workflow_file" "MODEL: github-models/deepseek/deepseek-v3-0324" "opencode review has a second reachable DeepSeek V3 fallback model"
+	assert_file_contains "$workflow_file" "Run OpenCode PR Review (DeepSeek R1)" "opencode review starts with DeepSeek R1"
+	assert_file_contains "$workflow_file" "MODEL: github-models/deepseek/deepseek-r1-0528" "opencode review starts with a reachable DeepSeek R1 reasoning model"
+	assert_file_contains "$workflow_file" "MODEL: github-models/deepseek/deepseek-v3-0324" "opencode review has a reachable DeepSeek V3 fallback model"
+	assert_file_contains "$workflow_file" "MODEL: github-models/openai/gpt-5" "opencode review still has a bounded GPT-5 fallback model"
 	assert_file_contains "$workflow_file" "Publish bounded OpenCode review comment" "opencode review workflow publishes the agent control comment for the approval gate"
 	assert_file_contains "$workflow_file" "statusCheckRollup" "opencode review workflow reads current-head GitHub Checks before approval"
 	assert_file_contains "$workflow_file" "OPENCODE_FAILED_CHECK_EVIDENCE_FILE" "opencode review workflow persists failed-check evidence across review and approval steps"
@@ -768,8 +769,8 @@ assert_opencode_review_uses_codegraph_and_gpt5_fallback() {
 	assert_file_contains "$workflow_file" '["FAILURE","TIMED_OUT","ACTION_REQUIRED","CANCELLED","STARTUP_FAILURE"]' "opencode review workflow treats failed check-run conclusions as request-changes blockers"
 	assert_file_contains "$workflow_file" '["FAILURE","ERROR"]' "opencode review workflow treats failed status contexts as request-changes blockers"
 	assert_file_not_contains "$workflow_file" "MODEL: github-models/gpt-4.1" "opencode review must not fall back to GPT-4.1"
-	assert_file_not_contains "$workflow_file" "MODEL: github-models/openai/gpt-5-chat" "opencode review must not use unavailable GitHub Models GPT-5 chat fallback"
-	assert_file_not_contains "$workflow_file" "MODEL: github-models/openai/gpt-5-mini" "opencode review must not use unavailable GitHub Models GPT-5 mini fallback"
+	assert_file_contains "$workflow_file" "github-models/openai/gpt-5-chat" "opencode review includes GitHub Models GPT-5 chat as a catalog fallback"
+	assert_file_contains "$workflow_file" "github-models/openai/gpt-5-mini" "opencode review includes GitHub Models GPT-5 mini as a catalog fallback"
 
 	assert_file_contains "$opencode_config" '"mcp"' "opencode config declares MCP servers"
 	assert_file_contains "$opencode_config" '"codegraph"' "opencode config declares the CodeGraph MCP server"
@@ -782,14 +783,15 @@ assert_opencode_review_uses_codegraph_and_gpt5_fallback() {
 	assert_file_contains "$opencode_config" '"serve"' "opencode config launches the CodeGraph MCP server"
 	assert_file_contains "$opencode_config" '"--mcp"' "opencode config launches CodeGraph in MCP mode"
 	assert_file_contains "$opencode_config" '"small_model": "github-models/deepseek/deepseek-v3-0324"' "opencode config uses a reachable DeepSeek V3 small model"
+	assert_file_contains "$opencode_config" '"model": "github-models/deepseek/deepseek-r1-0528"' "opencode config defaults review sessions to DeepSeek R1"
 	assert_file_contains "$opencode_config" '"openai/gpt-5"' "opencode config defines GitHub Models GPT-5 with full model id"
+	assert_file_contains "$opencode_config" '"openai/gpt-5-chat"' "opencode config defines GitHub Models GPT-5 chat fallback"
+	assert_file_contains "$opencode_config" '"openai/gpt-5-mini"' "opencode config defines GitHub Models GPT-5 mini fallback"
 	assert_file_contains "$opencode_config" '"deepseek/deepseek-r1-0528"' "opencode config defines DeepSeek R1 fallback"
 	assert_file_contains "$opencode_config" '"deepseek/deepseek-v3-0324"' "opencode config defines DeepSeek V3 fallback"
 	assert_file_contains "$opencode_config" '"context": 200000' "opencode config uses the GitHub Models GPT-5 200k context window"
 	assert_file_contains "$opencode_config" '"output": 100000' "opencode config uses the GitHub Models GPT-5 100k output window"
 	assert_file_not_contains "$opencode_config" "gpt-4.1" "opencode config must not define GPT-4.1 fallback"
-	assert_file_not_contains "$opencode_config" "gpt-5-chat" "opencode config must not define unavailable GPT-5 chat fallback"
-	assert_file_not_contains "$opencode_config" "gpt-5-mini" "opencode config must not define unavailable GPT-5 mini fallback"
 }
 
 assert_opencode_review_posts_suggested_diffs_inline() {
@@ -1380,7 +1382,7 @@ Model deepseek/deepseek-v3-0324 Vulnerabilities 1
 
 FAIL: strix workflow defaults PR Strix scans to GitHub Models GPT-5 (missing 'github.event.inputs.strix_llm || 'openai/gpt-5'')
 FAIL: strix workflow rejects unsupported model inputs (missing 'STRIX_LLM must select GitHub Models openai/gpt-5 or newer, direct OpenAI GPT-5.4 or newer, or an approved organization Vertex AI model')
-FAIL: opencode review tries GitHub Models GPT-5 first (missing 'MODEL: github-models/openai/gpt-5')
+FAIL: opencode review starts with DeepSeek R1 (missing 'MODEL: github-models/deepseek/deepseek-r1-0528')
 EOF
 	cat >"$control_json" <<'EOF'
 {"head_sha":"abc123","run_id":"42","run_attempt":"1","result":"REQUEST_CHANGES","reason":"Generic security concern","summary":"Generic speculative CI issues.","findings":[{"path":"scripts/ci/collect_failed_check_evidence.sh","line":15,"severity":"HIGH","title":"Generic finding","problem":"Speculative input validation issue unrelated to failed checks.","root_cause":"The review did not use the failed Strix evidence.","fix_direction":"Add generic validation.","regression_test_direction":"Add a generic test.","suggested_diff":"diff --git a/scripts/ci/collect_failed_check_evidence.sh b/scripts/ci/collect_failed_check_evidence.sh\n--- a/scripts/ci/collect_failed_check_evidence.sh\n+++ b/scripts/ci/collect_failed_check_evidence.sh\n@@ -1 +1 @@\n-old\n+new"}]}
@@ -1478,11 +1480,11 @@ Model deepseek/deepseek-v3-0324 Vulnerabilities 1
 
 FAIL: strix workflow defaults PR Strix scans to GitHub Models GPT-5 (missing 'github.event.inputs.strix_llm || 'openai/gpt-5'')
 FAIL: strix workflow rejects unsupported model inputs (missing 'STRIX_LLM must select GitHub Models openai/gpt-5 or newer, direct OpenAI GPT-5.4 or newer, or an approved organization Vertex AI model')
-FAIL: opencode review tries GitHub Models GPT-5 first (missing 'MODEL: github-models/openai/gpt-5')
+FAIL: opencode review starts with DeepSeek R1 (missing 'MODEL: github-models/deepseek/deepseek-r1-0528')
 EOF
 
 	cat >"$control_json" <<'EOF'
-{"head_sha":"abc123","run_id":"42","run_attempt":"1","result":"REQUEST_CHANGES","reason":"Strix Security Scan/strix failed","summary":"Strix Security Scan/strix failed in Self-test Strix gate script and reported github-models/openai/gpt-5 Authentication Bypass via X-Dev-User Header with Severity: CRITICAL at backend/app/auth.py:132-135 plus deepseek/deepseek-v3-0324 Frontend Security Issues: XSS, Hardcoded Credentials, and Insecure with Severity: HIGH.","findings":[{"path":".github/workflows/strix.yml","line":120,"severity":"HIGH","title":"Strix workflow default is not visible to trusted self-test","problem":"Strix Security Scan/strix failed in Self-test Strix gate script: strix workflow defaults PR Strix scans to GitHub Models GPT-5 (missing 'github.event.inputs.strix_llm || 'openai/gpt-5''); strix workflow rejects unsupported model inputs (missing 'STRIX_LLM must select GitHub Models openai/gpt-5 or newer, direct OpenAI GPT-5.4 or newer, or an approved organization Vertex AI model'); opencode review tries GitHub Models GPT-5 first (missing 'MODEL: github-models/openai/gpt-5'). The same failed Strix evidence includes github-models/openai/gpt-5 report Authentication Bypass via X-Dev-User Header, Severity: CRITICAL, /api/me, Method: GET, backend/app/auth.py:132-135.","root_cause":"The failed check evidence shows Self-test Strix gate script could not find github.event.inputs.strix_llm, STRIX_LLM must select, and MODEL: github-models/openai/gpt-5 in trusted-base files, and the model report identifies the backend auth fallback line.","fix_direction":"Update the workflow lines that provide the Strix model default and OpenCode model env so the trusted self-test can find those exact strings, then remove the unauthenticated X-Dev-User fallback at backend/app/auth.py:132-135.","regression_test_direction":"Keep the static self-test assertions for all three missing strings and add auth tests proving /api/me rejects forged X-Dev-User requests without signed auth.","suggested_diff":"diff --git a/.github/workflows/strix.yml b/.github/workflows/strix.yml\n--- a/.github/workflows/strix.yml\n+++ b/.github/workflows/strix.yml\n@@ -120 +120 @@\n-          STRIX_MODEL: old\n+          STRIX_MODEL: ${{ github.event.inputs.strix_llm || 'openai/gpt-5' }}"},{"path":"frontend/src/app/page.tsx","line":1,"severity":"HIGH","title":"Strix frontend model report must be reviewed separately","problem":"Strix Security Scan/strix failed with a separate deepseek/deepseek-v3-0324 report: Frontend Security Issues: XSS, Hardcoded Credentials, and Insecure, Severity: HIGH.","root_cause":"The failed Strix evidence contains a second model vulnerability report, so OpenCode must not collapse it into the first backend finding.","fix_direction":"Inspect the frontend source lines responsible for token storage, hardcoded credentials, dynamic error rendering, and missing CSP, then remove or harden each concrete line before approval.","regression_test_direction":"Add frontend tests covering safe token/session handling, output encoding, and security headers for the affected route.","suggested_diff":"diff --git a/frontend/src/app/page.tsx b/frontend/src/app/page.tsx\n--- a/frontend/src/app/page.tsx\n+++ b/frontend/src/app/page.tsx\n@@ -1 +1 @@\n-export default function Page() { return null }\n+export default function Page() { return null }"}]}
+{"head_sha":"abc123","run_id":"42","run_attempt":"1","result":"REQUEST_CHANGES","reason":"Strix Security Scan/strix failed","summary":"Strix Security Scan/strix failed in Self-test Strix gate script and reported github-models/openai/gpt-5 Authentication Bypass via X-Dev-User Header with Severity: CRITICAL at backend/app/auth.py:132-135 plus deepseek/deepseek-v3-0324 Frontend Security Issues: XSS, Hardcoded Credentials, and Insecure with Severity: HIGH.","findings":[{"path":".github/workflows/strix.yml","line":120,"severity":"HIGH","title":"Strix workflow default is not visible to trusted self-test","problem":"Strix Security Scan/strix failed in Self-test Strix gate script: strix workflow defaults PR Strix scans to GitHub Models GPT-5 (missing 'github.event.inputs.strix_llm || 'openai/gpt-5''); strix workflow rejects unsupported model inputs (missing 'STRIX_LLM must select GitHub Models openai/gpt-5 or newer, direct OpenAI GPT-5.4 or newer, or an approved organization Vertex AI model'); opencode review starts with DeepSeek R1 (missing 'MODEL: github-models/deepseek/deepseek-r1-0528'). The same failed Strix evidence includes github-models/openai/gpt-5 report Authentication Bypass via X-Dev-User Header, Severity: CRITICAL, /api/me, Method: GET, backend/app/auth.py:132-135.","root_cause":"The failed check evidence shows Self-test Strix gate script could not find github.event.inputs.strix_llm, STRIX_LLM must select, and MODEL: github-models/deepseek/deepseek-r1-0528 in trusted-base files, and the model report identifies the backend auth fallback line.","fix_direction":"Update the workflow lines that provide the Strix model default and OpenCode model env so the trusted self-test can find those exact strings, then remove the unauthenticated X-Dev-User fallback at backend/app/auth.py:132-135.","regression_test_direction":"Keep the static self-test assertions for all three missing strings and add auth tests proving /api/me rejects forged X-Dev-User requests without signed auth.","suggested_diff":"diff --git a/.github/workflows/strix.yml b/.github/workflows/strix.yml\n--- a/.github/workflows/strix.yml\n+++ b/.github/workflows/strix.yml\n@@ -120 +120 @@\n-          STRIX_MODEL: old\n+          STRIX_MODEL: ${{ github.event.inputs.strix_llm || 'openai/gpt-5' }}"},{"path":"frontend/src/app/page.tsx","line":1,"severity":"HIGH","title":"Strix frontend model report must be reviewed separately","problem":"Strix Security Scan/strix failed with a separate deepseek/deepseek-v3-0324 report: Frontend Security Issues: XSS, Hardcoded Credentials, and Insecure, Severity: HIGH.","root_cause":"The failed Strix evidence contains a second model vulnerability report, so OpenCode must not collapse it into the first backend finding.","fix_direction":"Inspect the frontend source lines responsible for token storage, hardcoded credentials, dynamic error rendering, and missing CSP, then remove or harden each concrete line before approval.","regression_test_direction":"Add frontend tests covering safe token/session handling, output encoding, and security headers for the affected route.","suggested_diff":"diff --git a/frontend/src/app/page.tsx b/frontend/src/app/page.tsx\n--- a/frontend/src/app/page.tsx\n+++ b/frontend/src/app/page.tsx\n@@ -1 +1 @@\n-export default function Page() { return null }\n+export default function Page() { return null }"}]}
 EOF
 	set +e
 	bash "$REPO_ROOT/scripts/ci/validate_opencode_failed_check_review.sh" \

From b4790f035d9aba3e58a99b390d2838d2ac1eaac2 Mon Sep 17 00:00:00 2001
From: Seongho Bae <me@seonghobae.me>
Date: Fri, 26 Jun 2026 13:19:36 +0900
Subject: [PATCH 2/2] Keep root OpenCode config base-compatible

---
 opencode.jsonc                      | 18 ------------------
 scripts/ci/test_strix_quick_gate.sh |  2 --
 2 files changed, 20 deletions(-)

diff --git a/opencode.jsonc b/opencode.jsonc
index 8adeee0e1..5d5382a88 100644
--- a/opencode.jsonc
+++ b/opencode.jsonc
@@ -107,24 +107,6 @@
             "output": 100000
           }
         },
-        "openai/gpt-5-chat": {
-          "name": "OpenAI GPT-5 Chat",
-          "tool_call": true,
-          "reasoning": true,
-          "limit": {
-            "context": 200000,
-            "output": 100000
-          }
-        },
-        "openai/gpt-5-mini": {
-          "name": "OpenAI GPT-5 Mini",
-          "tool_call": true,
-          "reasoning": true,
-          "limit": {
-            "context": 200000,
-            "output": 100000
-          }
-        },
         "deepseek/deepseek-r1-0528": {
           "name": "DeepSeek R1 0528",
           "tool_call": true,
diff --git a/scripts/ci/test_strix_quick_gate.sh b/scripts/ci/test_strix_quick_gate.sh
index e67f07c16..dd9cbd352 100755
--- a/scripts/ci/test_strix_quick_gate.sh
+++ b/scripts/ci/test_strix_quick_gate.sh
@@ -785,8 +785,6 @@ assert_opencode_review_uses_codegraph_and_gpt5_fallback() {
 	assert_file_contains "$opencode_config" '"small_model": "github-models/deepseek/deepseek-v3-0324"' "opencode config uses a reachable DeepSeek V3 small model"
 	assert_file_contains "$opencode_config" '"model": "github-models/deepseek/deepseek-r1-0528"' "opencode config defaults review sessions to DeepSeek R1"
 	assert_file_contains "$opencode_config" '"openai/gpt-5"' "opencode config defines GitHub Models GPT-5 with full model id"
-	assert_file_contains "$opencode_config" '"openai/gpt-5-chat"' "opencode config defines GitHub Models GPT-5 chat fallback"
-	assert_file_contains "$opencode_config" '"openai/gpt-5-mini"' "opencode config defines GitHub Models GPT-5 mini fallback"
 	assert_file_contains "$opencode_config" '"deepseek/deepseek-r1-0528"' "opencode config defines DeepSeek R1 fallback"
 	assert_file_contains "$opencode_config" '"deepseek/deepseek-v3-0324"' "opencode config defines DeepSeek V3 fallback"
 	assert_file_contains "$opencode_config" '"context": 200000' "opencode config uses the GitHub Models GPT-5 200k context window"