GibilGibil

PR Review with Test Results

Review pull requests with actual test results, not just static analysis

An agent checks out a PR branch, builds the project, runs the full test suite, and reviews with real pass/fail data — not just static analysis.

Workflow

# Forge a server with the repo
gibil create --name review --repo github.com/you/project --ttl 30 --json

# Fetch and checkout the PR branch
gibil run review "cd /root/project && git fetch origin pull/42/head:pr-42 && git checkout pr-42" --json

# Install deps and run the full suite
gibil run review "cd /root/project && pnpm install && pnpm test" --json
# → {"stdout": "...", "stderr": "", "exit_code": 0}

# Agent posts review with actual test results
gibil destroy review

The agent's review includes "I ran the tests and they pass" — not "the code looks correct."

Variation: dependency upgrade impact

Forge a fleet and test multiple upgrades in parallel. Report which ones break the build.

gibil create --name upgrade --fleet 3 --repo github.com/you/project --ttl 30 --json

gibil run upgrade-1-abc "cd /root/project && npm install react@19 && pnpm test" --json
gibil run upgrade-2-abc "cd /root/project && npm install next@15 && pnpm test" --json
gibil run upgrade-3-abc "cd /root/project && npm install typescript@6 && pnpm test" --json

gibil destroy --all

Each upgrade runs on a clean server — no cross-contamination between upgrades.

Why gibil

  • Evidence-based reviews — the agent ran the code, not just read it
  • Clean server — test results reflect the PR in isolation, not leftover state
  • Fleet for comparisons — test N upgrades or N consumer projects in parallel

Set GITHUB_TOKEN as an environment variable to allow the agent to push branches and open PRs directly from the server. See Remote PR Workflow.

Next steps

On this page