claude-teammate 0.1.41 → 0.1.43

package/README.md

@@ -41,7 +41,9 @@ Assign it tickets. It ships.
 - ✅ You want Jira tickets to **turn into merged PRs** without manual handoff
 - ✅ You want **PR review comments acted on**, not just acknowledged
 - ✅ You need a reviewer that **always shows up** and leaves structured feedback
+- ✅ You want **automated PR review** without adding another review subscription on top of Claude
 - ✅ You want an AI that **remembers** - which repo, what's shipped, what breaks, what to avoid
+- ✅ You want the agent running in **its own working environment** so it can test code and verify real outputs
 - ✅ You want this running on **your own infrastructure**
 
 <br/>
@@ -72,10 +74,22 @@ Add it as a reviewer and it will show up - every time, with real feedback, never
 
 <br/>
 
+## Review Human PRs Too
+
+Claude Teammate is not limited to tickets it implemented itself. Add it as a reviewer on a human-authored pull request and it will read the diff, inspect the repository, and leave structured review feedback. This can be an alternative to paid tools like CodeRabbit or GitHub Copilot code review.
+
+<p align="center">
+  <img src="assets/PR_review.webp" alt="Claude Teammate reviewing a human pull request" width="760" style="border: 1px solid #d1d5db;" />
+</p>
+
+<br/>
+
 ## It Remembers
 
 Every epic has its own memory. The bot picks up exactly where it left off - no matter how much time has passed.
 
+That memory is built from more than the original Jira description. Claude Teammate updates it from planning, implementation, review comments, and recurring human feedback so future tickets inherit the durable parts of the work instead of starting from scratch.
+
 ```
 memory/
 └── {domain}/
@@ -87,7 +101,9 @@ memory/
 |---|---|
 | `jira_key` / `jira_url` | Canonical epic identity |
 | `repos` | Shared target GitHub repositories for the epic |
-| `summary` / `guardrails` | Reusable lessons, build quirks, verification workflows, and what-not-to-do notes for the epic |
+| `facts` / `guardrails` | Reusable lessons, build quirks, verification workflows, recurring human feedback, and what-not-to-do notes for the epic |
+
+The memory stays compact on purpose. It is cleaned up, deduplicated, and summarized over time so later tickets can reuse the important context without dragging an ever-growing prompt window behind them.
 
 <br/>
 
@@ -162,6 +178,9 @@ Yes. Each epic is scoped to its own Jira domain and workspace, and can map to on
 **Does it understand the codebase or just generate generic code?**
 It reads the actual repository before writing anything - structure, conventions, existing patterns. The plan it proposes is specific to your code, not a template.
 
+**Can it test code and verify UI changes?**
+Yes. It works in its own cloned working environment, can run the repo's test and build commands, and can use Playwright to take screenshots and verify frontend changes instead of relying only on the diff.
+
 **Can I use Claude skills or MCP, and does it respect `CLAUDE.md`?**
 Yes. Claude skills, configured MCP servers, and repo instructions in `CLAUDE.md` are available and respected just like they are in a normal Claude-driven workflow. Playwright MCP is built in and enabled by default so challenging frontend tasks can be verified in-browser.
 

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "claude-teammate",
-  "version": "0.1.41",
+  "version": "0.1.43",
   "description": "CLI bootstrapper for Claude Teammate.",
   "license": "MIT",
   "type": "module",

package/src/claude.js

@@ -1037,7 +1037,7 @@ function buildGitHubPRReviewSystemPrompt() {
     "Skip trivial style-only issues unless they introduce a real risk.",
     "Classify each suggestion as Minor (non-blocking) or Major (should be addressed before merge).",
     "Where possible, produce a committable suggestion with the corrected code.",
-    "Return a summary paragraph describing what the PR does, followed by a summary of all suggestions.",
+    "Return a summary with the following structure: Short overview (1–3 sentences), Key changes (bullet points), Summary of all suggestions (1-2 sentences).",
     "Return only structured output matching the provided schema."
   ].join(" ");
 }
@@ -1200,13 +1200,16 @@ Pull request number: ${input.pr?.number ?? ""}
 Pull request title:
 ${input.pr?.title ?? ""}
 
+Pull request description:
+${input.pr?.body ?? "(empty)"}
+
 Repository path:
 ${input.repoPath}
 
 Unified diff:
 ${input.diff || "(empty)"}
 
-Review the diff carefully. Read related code files with Read, Grep, or Glob only when the diff alone is insufficient to judge correctness. Return a summary describing the PR and its suggestions.`;
+Return a concise PR summary in the schema summary field, plus any review suggestions in the schema suggestions field.`;
 }
 
 function buildEpicMemoryCleanupUserPrompt(input) {

package/src/commands/worker.js

@@ -377,28 +377,13 @@ export async function runWorkerCommand({ projectRoot }) {
         }
 
         try {
-          const repoPath = await ensureReviewRepo(pr.repoUrl, runtimePaths.reviewReposDir);
-          const prDetail = await github.fetchPullRequest(pr.repoUrl, pr.number);
-          await checkoutPullRequestBranch(repoPath, prDetail.headRef, prDetail.headRepoCloneUrl);
-
-          const diff = await github.fetchPullRequestDiff(pr.repoUrl, pr.number);
-          const result = await runClaudePrReview({
-            diff,
-            repoPath,
-            pr: prDetail,
+          const { prDetail, result } = await processReviewPullRequest({
+            pr,
+            github,
+            reviewReposDir: runtimePaths.reviewReposDir,
             timeoutMs: parseOptionalInt(values.CLAUDE_TIMEOUT_MS)
           });
 
-          await github.createPullRequestReview(
-            pr.repoUrl,
-            pr.number,
-            result.summary,
-            result.suggestions,
-            "COMMENT"
-          );
-
-          await github.addLabelsToPullRequest(pr.repoUrl, pr.number, ["AI-reviewed"]);
-
           processedPrs.push({
             repoUrl: pr.repoUrl,
             pullRequestNumber: String(pr.number),
@@ -455,6 +440,45 @@ export async function runWorkerCommand({ projectRoot }) {
   }, POLL_INTERVAL_MS);
 }
 
+export async function processReviewPullRequest({
+  pr,
+  github,
+  reviewReposDir,
+  timeoutMs,
+  ensureReviewRepoFn = ensureReviewRepo,
+  fetchPullRequestFn = (repoUrl, prNumber) => github.fetchPullRequest(repoUrl, prNumber),
+  checkoutPullRequestBranchFn = checkoutPullRequestBranch,
+  fetchPullRequestDiffFn = (repoUrl, prNumber) => github.fetchPullRequestDiff(repoUrl, prNumber),
+  runClaudePrReviewFn = runClaudePrReview,
+  createPullRequestReviewFn = (repoUrl, prNumber, body, suggestions, event) =>
+    github.createPullRequestReview(repoUrl, prNumber, body, suggestions, event)
+}) {
+  await github.addLabelsToPullRequest(pr.repoUrl, pr.number, ["AI-reviewed"]);
+  await github.postIssueComment(pr.repoUrl, pr.number, "I am reviewing the PR.");
+
+  const repoPath = await ensureReviewRepoFn(pr.repoUrl, reviewReposDir);
+  const prDetail = await fetchPullRequestFn(pr.repoUrl, pr.number);
+  await checkoutPullRequestBranchFn(repoPath, prDetail.headRef, prDetail.headRepoCloneUrl);
+
+  const diff = await fetchPullRequestDiffFn(pr.repoUrl, pr.number);
+  const result = await runClaudePrReviewFn({
+    diff,
+    repoPath,
+    pr: prDetail,
+    timeoutMs
+  });
+
+  await createPullRequestReviewFn(
+    pr.repoUrl,
+    pr.number,
+    result.summary,
+    result.suggestions,
+    "COMMENT"
+  );
+
+  return { prDetail, result };
+}
+
 async function processJiraIssue({ issue, jira, github, botUser, config, projectRoot, runtimePaths, logger }) {
   const detail = await jira.fetchIssueDetails(issue.key);