docs(archive): Preserve discussion for PR #5514

docs/pr-discussions/PR-5514-docs-research-add-drift-report-on-agent-paralysis.md

@@ -0,0 +1,73 @@
+---
+pr_number: 5514
+title: "docs(research): Add drift report on agent paralysis"
+author: "AceHack"
+state: "MERGED"
+created_at: "2026-05-27T17:05:10Z"
+merged_at: "2026-05-27T17:18:08Z"
+closed_at: "2026-05-27T17:18:09Z"
+head_ref: "lior/agent-paralysis-drift-report-2026-05-27"
+base_ref: "main"
+archived_at: "2026-05-27T17:44:30Z"
+archive_tool: "tools/pr-preservation/archive-pr.ts"
+---
+
+# PR #5514: docs(research): Add drift report on agent paralysis
+
+## PR description
+
+This PR contains a drift report detailing the paralysis of agents Otto, Riven, and Kiro, as well as Lior's self-correction on PR noise. This PR addresses the issues from PR #5496.
+
+## Reviews
+
+### COMMENTED — @copilot-pull-request-reviewer (2026-05-27T17:08:05Z)
+
+## Pull request overview
+
+Adds a new research drift report documenting a multi-agent “paralysis” incident in the Zeta factory (Otto/Kiro/Riven) and records recommended mitigations around worktree/clone isolation and PR-noise reduction.
+
+**Changes:**
+- Introduces a structured drift report with executive summary, per-agent findings, and recommendations.
+- Captures concrete operational observations (dirty/contested worktrees, volatile `/tmp` worktree path) and proposed remediation steps.
+
+### COMMENTED — @AceHack (2026-05-27T17:14:28Z)
+
+_(no body)_
+
+### COMMENTED — @AceHack (2026-05-27T17:14:38Z)
+
+_(no body)_
+
+### COMMENTED — @copilot-pull-request-reviewer (2026-05-27T17:15:45Z)
+
+## Pull request overview
+
+Copilot reviewed 1 out of 1 changed files in this pull request and generated no new comments.
+
+## Review threads
+
+### Thread 1: docs/research/agent-paralysis-drift-report-2026-05-27.md:18 (resolved)
+
+**@copilot-pull-request-reviewer** (2026-05-27T17:08:04Z):
+
+P1: The reference to “isolated worktree protocol (B-0751)” is not currently resolvable in-tree (there is no backlog row/document with `id: B-0751`). This makes the citation hard to verify; consider citing the concrete artefact that exists (e.g., PR #5019 per-agent isolated clones) and/or a backlog row that exists (e.g., B-0750 worktree hygiene), and align the wording with what B-0751 is described as elsewhere (“per-agent isolated clones architecture”).
+
+**@AceHack** (2026-05-27T17:14:28Z):
+
+Fixed in 67aa9848f — replaced the B-0751 reference at line 18 with PR #5019 (closed proposal) + B-0750 (worktree-hygiene row in tree) + agent-worktree-hygiene rule path. All three artifacts are resolvable in-tree.
+
+### Thread 2: docs/research/agent-paralysis-drift-report-2026-05-27.md:39 (resolved)
+
+**@copilot-pull-request-reviewer** (2026-05-27T17:08:05Z):
+
+P1: “per-agent isolated clone/worktree architecture (B-0751)” appears to cite a backlog row that isn’t present in the repo (`id: B-0751` not found). To keep this recommendation actionable, cite the existing source of truth (PR #5019 and/or the relevant backlog row such as B-0750) or add a linkable reference that exists in-tree.
+
+**@AceHack** (2026-05-27T17:14:38Z):
+
+Fixed in 67aa9848f — replaced the B-0751 reference at line 38 with PR #5019 (closed design proposal) + B-0750 (worktree-hygiene row that exists in tree). Both are linkable + verifiable.
+
+## General comments
+
+### @chatgpt-codex-connector (2026-05-27T17:05:14Z)
+
+You have reached your Codex usage limits for code reviews. You can see your limits in the [Codex usage dashboard](https://chatgpt.com/codex/cloud/settings/usage).