open-pr

Take a cloudflare/agents GitHub issue plus any repro findings and one-shot a fix PR — branch, change, test, push, and open the PR linked to the issue.

npx skills add https://github.com/cloudflare/agents --skill open-pr

The current user message contains an <agent-think-run> envelope with repository, issue, instruction, requested-by, and (when available) trigger-comment-id. Use those values exactly. Never infer or substitute another target from examples, workspace contents, GitHub searches, or concurrent issues. If the envelope or a required field is absent, stop without cloning/editing/pushing/posting and return a structured skipped result. When trigger-comment-id is present, your first container action is the liveness reaction:

gh api repos/<repository>/issues/comments/<trigger-comment-id>/reactions \
  -f content=rocket

Produce a focused fix PR, authored as yourself — the agent-think GitHub App. Do not impersonate any user.

The instruction is the free-form text the user typed after @agent-think (it may be empty). Treat it as a direct instruction — e.g. constraints on the fix, a preferred approach, or a pointer to the suspect area — and weight it highly, but stay within the scope of the issue.

All gh, git, npm, curl, and wrangler commands must run on the container backend (bash({ command, backend: "container" })) — the shell backend has no real binaries or network. gh is already authenticated as the app; use it directly (no token handling). Work under /workspace; put long logs and disposable scratch files in container-local /temp.

0. Clone the repo

Clone the target repo directly under /workspace using its repository name (cloudflare/agents/workspace/agents):

REPO_DIR="/workspace/$(basename <repo>)"
if [ ! -d "$REPO_DIR/.git" ]; then
  git clone https://github.com/<repo>.git "$REPO_DIR"
fi
cd "$REPO_DIR"

1. Gather everything known

gh issue view <issueNumber> --repo <repo> --json title,body,labels,author,comments

Read the body and every comment. In particular look for a comment from the prior repro run (@agent-think repro): it may contain a minimal reproduction, a live URL, observed-vs-expected behavior, and a root-cause hypothesis pointing at a file/line. Use it as your starting point — do not re-derive what is already known.

Decide if this is fixable in one shot. If the issue is a feature request needing design, is too vague, spans many subsystems, or you cannot locate a confident root cause, stop: return prOpened: false, skipped: true, and a summary explaining why. Post a brief, polite comment saying the pr-agent is skipping it and what additional detail would help; begin it with Requested by @<requestedBy> when the run envelope has a requester.

2. Locate the root cause

With the repo cloned at $REPO_DIR, read the relevant code in packages/agents, packages/think, etc. Confirm the hypothesis (or form your own) by reading the actual implementation. Identify the smallest change that fixes the reported behavior.

3. Branch and set commit identity

Commit as the agent-think app itself:

git config user.name "agent-think[bot]"
git config user.email "agent-think[bot]@users.noreply.github.com"
BRANCH="fix/issue-<issueNumber>-$(date +%s)"
git checkout -b "$BRANCH"

4. Make the fix

  • Smallest correct change. Fewest files. Match the existing code style.
  • Add or update a test that fails before the fix and passes after, when the area is testable.
  • Add a changeset if the repo uses them (.changeset/) for a user-facing fix: create a markdown file following the existing format in .changeset/.
  • Update examples and docs when the change affects behavior they show: if an examples/* app or a docs//README section demonstrates the code you touched, keep it truthful in the same PR.

5. Verify

Install and run the affected package's checks (monorepo uses pnpm + Nx):

# NOISY commands (installs, builds, test suites) MUST be redirected to a
# container-local /temp file and tailed — streaming megabytes of live output
# through the session can kill it irrecoverably:
mkdir -p /temp
CI=1 pnpm install --frozen-lockfile --reporter=append-only \
  > /temp/install.log 2>&1 || (tail -40 /temp/install.log; false)
tail -20 /temp/install.log
# Prefer scoped/affected runs; fall back to package scripts.
pnpm -w exec oxfmt --check . || pnpm -w exec oxfmt --write .
pnpm -w exec oxlint . || true
# Run the relevant package's typecheck + tests, redirected the same way:
pnpm --filter <package> typecheck > /temp/typecheck.log 2>&1; tail -30 /temp/typecheck.log
pnpm --filter <package> test > /temp/test.log 2>&1; tail -40 /temp/test.log

(/temp is outside the /workspace mount and is not synchronized.)

Record whether tests passed in testsPassed. If you cannot make tests pass and the failure is your change's fault, fix it; if tests are unrelated/flaky, note that in the PR body. Do not open a PR whose own new test fails.

6. Deploy a live demo of the PR

Every PR ships with a temporary deployment demoing the change so reviewers click a link and see the fixed behavior instead of imagining it.

  1. Build the fixed package(s) in the repo and pack them:
pnpm --filter <package> build
(cd packages/<package> && npm pack --pack-destination /workspace)   # -> /workspace/<package>-x.y.z.tgz
  1. Build a minimal demo app in /workspace/demo-<issueNumber> that exercises the fixed path. Follow the activated reproduce skill's "Minimal frontend (required)" recipe — same 7 files — but install the packed tarball so the demo runs YOUR fix:
npm install /workspace/<package>-x.y.z.tgz

If the repro run left a repro/issue-<issueNumber> branch, start from that project instead and just swap the dependency to the packed build — the same UI then demos broken-before / fixed-after.

  1. Deploy and verify the fix is actually observable in the page:
npm run deploy      # vite build && wrangler deploy --temporary
curl -sS -i "<demoUrl>/" | head -5

Capture the https://...workers.dev URL as demoUrl. Ignore the claim URL the deploy prints — it never goes in a PR.

Use judgment: changes with no runtime surface (docs-only, types-only, CI) skip the demo — say so in the PR body ("no runtime surface to demo") rather than deploying something meaningless.

7. Commit, push, open the PR

git status --short
git add -A
git commit -m "fix: <concise description> (#<issueNumber>)"
git push -u origin "$BRANCH"

gh pr create --repo <repo> \
  --base main \
  --head "$BRANCH" \
  --title "fix: <concise description> (#<issueNumber>)" \
  --body-file /temp/pr-body.md

Write the PR body outside the checkout at /temp/pr-body.md. It must include:

  • Requested by @<requestedBy> near the top, using the exact sanitized requested-by mention from the run envelope (omit only when it is unknown).
  • Closes #<issueNumber> so the issue auto-links.
  • What was wrong (root cause, citing the file/line).
  • What changed and why this is the minimal fix.
  • Testing: what you added/ran and the result.
  • Demo: phrase it exactly like "Demo URL (expires after 60 mins): " plus one line of click instructions and a note that it runs the packed build from this branch. If skipped, one line on why.
  • A link back to the repro-agent's reproduction (comment and/or the repro/issue-<issueNumber> branch) if one exists.
  • A "🤖 generated by the pr-agent — please review carefully" footer.

Capture the PR URL and branch name.

Do not post an acknowledgement or "opened a PR" comment on the issue. The PR links back to the issue via Closes #<issueNumber>. The only case where you comment is the skip path in the Gather everything known step (no PR exists to show).

8. Return the structured result

Return exactly:

  • prOpened (boolean)
  • skipped (boolean)
  • summary (string — one or two sentences)
  • prUrl (string, optional)
  • branch (string, optional)
  • testsPassed (boolean, optional)
  • demoUrl (string, optional)

More skills from cloudflare