ci: switch gadfly review to the reusable workflow (curated swarm, 5 lenses) #6

Merged
steve merged 2 commits from ci/gadfly-reusable into main 2026-06-28 02:48:28 +00:00
2 changed files with 26 additions and 53 deletions
+21 -50
View File
@@ -1,12 +1,8 @@
# Gadfly — agentic adversarial PR reviewer (https://gitea.stevedudenhoeffer.com/steve/gadfly). # Gadfly adversarial review — subscribes to steve/gadfly's reusable workflow and
# # INHERITS its default swarm. This stub holds only the triggers, the actor gate,
# Runs the published Gadfly image (pinned to an immutable :sha- tag — act_runner # secret forwarding, and the allow-list; the swarm config (models, lenses,
# caches :latest, and this build is what carries foreman provider-type support) # concurrency, timeouts) lives centrally in gadfly's review-reusable.yml so it is
# as a specialist swarm and posts # tuned in ONE place. Advisory only — never blocks a merge.
# ONE consolidated review comment as gitea-actions. Advisory only — never blocks a
# merge. This reviews majordomo PRs with 6 ollama-cloud models (3-lens suite).
# Gadfly is a simple system — findings are advisory; always double-check before
# acting.
name: Adversarial Review (Gadfly) name: Adversarial Review (Gadfly)
@@ -33,50 +29,25 @@ concurrency:
jobs: jobs:
review: review:
# Security: only trusted users may trigger a secret-bearing run via a PR # Security: only trusted users may trigger a secret-bearing run via a PR
# comment (pull_request + workflow_dispatch are already trusted). Mirrors # comment (pull_request + workflow_dispatch are already trusted). Mirrors the
# GADFLY_ALLOWED_USERS, the in-container belt-and-suspenders check. # allowed_users input below (the in-container belt-and-suspenders check) — both
# lists must stay in sync; a workflow if: can't read a workflow_call input.
if: >- if: >-
github.event_name != 'issue_comment' github.event_name != 'issue_comment'
|| (github.event.issue.pull_request || (github.event.issue.pull_request
&& (github.actor == 'steve' && (github.actor == 'steve'
|| github.actor == 'fizi' || github.actor == 'fizi'
|| github.actor == 'dazed')) || github.actor == 'dazed'))
runs-on: ubuntu-latest # Pinned to an immutable gadfly commit (not @main): a push to gadfly can't
# Fleet: 6 ollama-cloud models (lens fan-out), no local Macs. (Trimmed the # silently change the code that runs with our forwarded secrets.
# weakest reviewers by grade — m5/qwen3.6, gemma4, gpt-oss, kimi-k2.7 — plus uses: steve/gadfly/.gitea/workflows/review-reusable.yml@b02b11d69139843665da4cdbf776bc0b3583490d
# the earlier M1 drop.) Plenty of headroom for the cloud lanes. # Least privilege: forward only the review secrets (not `secrets: inherit`,
timeout-minutes: 45 # which would expose every repo secret). GITEA_TOKEN is the automatic token.
steps: secrets:
- uses: docker://gitea.stevedudenhoeffer.com/steve/gadfly:sha-d7f364d OLLAMA_CLOUD_API_KEY: ${{ secrets.OLLAMA_CLOUD_API_KEY }}
env: CLAUDE_CODE_OAUTH_TOKEN: ${{ secrets.CLAUDE_CODE_OAUTH_TOKEN }}
GITEA_API: ${{ github.server_url }}/api/v1/repos/${{ github.repository }} GADFLY_FINDINGS_URL: ${{ secrets.GADFLY_FINDINGS_URL }}
GITEA_TOKEN: ${{ secrets.GITEA_TOKEN }} GADFLY_FINDINGS_TOKEN: ${{ secrets.GADFLY_FINDINGS_TOKEN }}
OLLAMA_CLOUD_API_KEY: ${{ secrets.OLLAMA_CLOUD_API_KEY }} with:
# Cloud-only fleet (no local Macs). Cloud concurrency lives in the # Consumer-specific allow-list; everything else is inherited.
# LENSES: models run a few at a time (ollama-cloud=3) with their 3 lenses allowed_users: "steve,fizi,dazed"
# concurrent (LENS ollama-cloud=3) so comments land sooner.
GADFLY_MODELS: "minimax-m3:cloud,glm-5.2:cloud,glm-5.1:cloud,deepseek-v4-pro:cloud,nemotron-3-super:cloud,qwen3-coder:480b-cloud"
GADFLY_PROVIDER_CONCURRENCY: "ollama-cloud=3"
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").
GADFLY_SPECIALISTS: ${{ vars.GADFLY_SPECIALISTS || 'security,correctness,error-handling' }}
# Per-lens deadline + bounded steps to keep each reviewer's run sane.
GADFLY_TIMEOUT_SECS: "600"
GADFLY_MAX_STEPS: "14"
# Allow-list for the comment trigger (mirrors the job-level if: guard).
GADFLY_ALLOWED_USERS: "steve,fizi,dazed"
# --- findings telemetry: POST runs + findings to the gadfly-reports store ---
# Advisory & off unless GADFLY_FINDINGS_URL is set; failures only log to
# stderr and never affect the review. GADFLY_REPO / GADFLY_PR are derived
# in-container; the URL + token are user-scope secrets.
GADFLY_FINDINGS_URL: ${{ secrets.GADFLY_FINDINGS_URL }}
GADFLY_FINDINGS_TOKEN: ${{ secrets.GADFLY_FINDINGS_TOKEN }}
# --- event context (leave as-is) ---
EVENT_NAME: ${{ github.event_name }}
PR: ${{ github.event.pull_request.number || github.event.issue.number || github.event.inputs.pr_number }}
PR_BRANCH: ${{ github.head_ref }}
IS_DRAFT: ${{ github.event.pull_request.draft }}
COMMENT_BODY: ${{ github.event.comment.body }}
COMMENT_ID: ${{ github.event.comment.id }}
ACTOR: ${{ github.actor }}
+5 -3
View File
@@ -142,9 +142,11 @@ Ship work through PRs and let Gadfly review it before merge:
- **Push to a PR, never straight to `main`.** Branch, push, open a PR. - **Push to a PR, never straight to `main`.** Branch, push, open a PR.
`.gitea/workflows/adversarial-review.yml` runs Gadfly (the standalone `.gitea/workflows/adversarial-review.yml` runs Gadfly (the standalone
agentic adversarial reviewer) — a fleet of 6 ollama-cloud models, each agentic adversarial reviewer) by subscribing to gadfly's reusable workflow
running the 3-lens suite (security, correctness, error-handling). Advisory and inheriting its default swarm — 3 cloud models + the Claude Code engine
only; it never blocks the merge. (sonnet/opus/opus:max), each running the 5-lens suite (security, correctness,
maintainability, performance, error-handling). The swarm is tuned centrally
in gadfly, not here. Advisory only; it never blocks the merge.
- **Wait for Gadfly to finish, then read its output.** Don't merge while the - **Wait for Gadfly to finish, then read its output.** Don't merge while the
review is still running. Each model posts one consolidated comment; weigh review is still running. Each model posts one consolidated comment; weigh
every finding on its merits and fix the real ones (Gadfly is a simple every finding on its merits and fix the real ones (Gadfly is a simple