ci(gadfly): add M1/M5 Macs back to the reviewer pool (full fleet)

Re-adds the local Macs (m1/qwen3:14b, m5/qwen3.6:35b-mlx) via their foreman endpoints alongside the 3 cloud models. Cloud keeps lens fan-out (ollama-cloud=1 model + lens=3); each Mac runs one model with lenses serial (foreman serializes anyway); all provider lanes parallel. Bumps the job timeout 30->90m for the slow local lanes. With findings telemetry now on, gadfly-reports can quantify whether the Macs earn their keep.

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>

.gitea/workflows/adversarial-review.yml

@@ -41,25 +41,33 @@ jobs:
               || github.actor == 'fizi'
               || github.actor == 'dazed'))
     runs-on: ubuntu-latest
-    # 3 cloud models, all concurrent, 3-lens suite. ~12 min typical.
-    timeout-minutes: 30
+    # Full fleet: 3 cloud (lens fan-out) + M1/M5 Macs via foreman. The slow local
+    # lanes dominate wall time, so allow plenty of headroom.
+    timeout-minutes: 90
     steps:
       - uses: docker://gitea.stevedudenhoeffer.com/steve/gadfly:sha-d7f364d
         env:
           GITEA_API: ${{ github.server_url }}/api/v1/repos/${{ github.repository }}
           GITEA_TOKEN: ${{ secrets.GITEA_TOKEN }}
           OLLAMA_CLOUD_API_KEY: ${{ secrets.OLLAMA_CLOUD_API_KEY }}
-          # executus uses CLOUD MODELS ONLY. The local Macs (m1/m5) were dropped:
-          # on a P2-review measurement they took 26–29 min (with lens timeouts)
-          # and contributed ZERO real findings — the two cloud models found every
-          # genuine bug in 6–12 min. Cloud-only is faster AND higher-signal.
-          # 3 cloud models. Concurrency now lives in the LENSES, not the models:
-          # one model runs at a time (PROVIDER_CONCURRENCY=1) with its 3 lenses
-          # concurrent (PROVIDER_LENS_CONCURRENCY=3). So the first model's
-          # comment lands sooner and each model finishes a bit faster, at the
-          # cost of the other two models' comments arriving in series after it.
-          GADFLY_MODELS: "minimax-m3:cloud,deepseek-v4-flash:cloud,glm-5.2:cloud"
-          GADFLY_PROVIDER_CONCURRENCY: "ollama-cloud=1"
+          # Local Macs, reached through their foreman queues (native Ollama on the
+          # wire). GADFLY_ENDPOINT_M1 registers provider "m1", _M5 registers "m5",
+          # each a foreman-preset Ollama client at the secret's URL, of the form:
+          #   foreman|https://<foreman-host>|<token>
+          # Needs an image with foreman provider-type support (this one). If a Mac
+          # is offline that model's comment shows an error and the others still post.
+          # (Gitea secrets aren't auto-exposed — map each explicitly.)
+          GADFLY_ENDPOINT_M1: ${{ secrets.GADFLY_ENDPOINT_M1 }}
+          GADFLY_ENDPOINT_M5: ${{ secrets.GADFLY_ENDPOINT_M5 }}
+          # Full fleet: 3 cloud + M1 Pro + M5 Max. The Macs are back so the
+          # gadfly-reports scoreboard can quantify whether they earn their keep
+          # (they previously took 26–29 min for ZERO real findings — now measured).
+          # Cloud concurrency lives in the LENSES: one cloud model at a time
+          # (ollama-cloud=1) with its 3 lenses concurrent (LENS ollama-cloud=3) so
+          # its comment lands sooner; each Mac runs one model, lenses serial (its
+          # foreman queue serializes anyway). All three provider lanes run parallel.
+          GADFLY_MODELS: "minimax-m3:cloud,deepseek-v4-flash:cloud,glm-5.2:cloud,m1/qwen3:14b,m5/qwen3.6:35b-mlx"
+          GADFLY_PROVIDER_CONCURRENCY: "ollama-cloud=1,m1=1,m5=1"
           GADFLY_PROVIDER_LENS_CONCURRENCY: "ollama-cloud=3"
           # Default => the 3-lens suite (security, correctness, error-handling).
           # Set the repo var GADFLY_SPECIALISTS to override (csv / "all" / "auto").