<?xml version="1.0" encoding="UTF-8"?><rss version="2.0"><channel><title>Windtask</title><description>Field notes on coding agents, leftover cloud infrastructure, and the bills that outlive the chat.</description><link>https://windtask.com/</link><language>en-us</language><item><title>AI Agent Identity: Why Shared Employee Keys Explode Blast Radius</title><link>https://windtask.com/articles/ai-agent-identity-permissions-mintive/</link><guid isPermaLink="true">https://windtask.com/articles/ai-agent-identity-permissions-mintive/</guid><description>Treat each AI agent as its own principal—scoped tokens, delegated permissions, one-agent revoke. Compare Mintive, Okta Govern AI Agent Identity, and WorkOS Agent Auth; draw the boundary vs an MCP gateway.</description><pubDate>Wed, 04 Nov 2026 00:00:00 GMT</pubDate><category>Security</category><category>ai-agents</category><category>agent-identity</category><category>mintive</category><category>okta</category><category>workos</category><category>mcp</category><category>iam</category><category>permissions</category><category>zero-trust</category></item><item><title>Claude Self-Assessment Privacy: A Decision Guide by Account Type</title><link>https://windtask.com/articles/claude-self-assessment-snitch-privacy/</link><guid isPermaLink="true">https://windtask.com/articles/claude-self-assessment-snitch-privacy/</guid><description>Stop guessing from viral &apos;digital snitch&apos; posts. Decide by account type whether self-assessment runs, what retention applies, how Cyber Verification handles research false positives, and why Claude Code telemetry is not server-side assessment.</description><pubDate>Wed, 04 Nov 2026 00:00:00 GMT</pubDate><category>Privacy</category><category>claude</category><category>self-assessment</category><category>privacy</category><category>zdr</category><category>cyber-verification</category><category>retention</category><category>claude-code</category></item><item><title>Coding-Agent Hooks Hygiene: Which File Still Owns the Entry</title><link>https://windtask.com/articles/coding-agent-hooks-hygiene-settings-json/</link><guid isPermaLink="true">https://windtask.com/articles/coding-agent-hooks-hygiene-settings-json/</guid><description>Decide which file owns a third-party hook, which official unload removes it, and which managed hooks a user disable cannot turn off after an orchestrator or desktop installer rewrites Claude, Codex, Cursor, or Copilot config.</description><pubDate>Wed, 04 Nov 2026 00:00:00 GMT</pubDate><category>Developer Tools</category><category>coding-agents</category><category>hooks</category><category>claude-code</category><category>codex</category><category>cursor</category><category>settings-json</category><category>agent-orchestrators</category><category>postman</category></item><item><title>Parallel Coding-Agent Conflicts: Intent Claims vs Contracts vs Ownership</title><link>https://windtask.com/articles/foremerge-parallel-agent-intent-conflicts/</link><guid isPermaLink="true">https://windtask.com/articles/foremerge-parallel-agent-intent-conflicts/</guid><description>Decide which parallel-agent conflict class you are solving before spinning N worktrees. Compare deterministic intent claims (scope+operation), contract-first merges, ownership orchestration, and LLM mutual review—with Foremerge as a worked case, not a launch post.</description><pubDate>Wed, 04 Nov 2026 00:00:00 GMT</pubDate><category>Developer Tools</category><category>foremerge</category><category>parallel-agents</category><category>coding-agents</category><category>claude-code</category><category>codex</category><category>cursor</category><category>intent-conflicts</category><category>multi-agent</category></item><item><title>Playwright Chromium Lag for Coding Agents: A Recurring Trust Problem, Not a One-Off CVE</title><link>https://windtask.com/articles/playwright-chromium-lag-coding-agents/</link><guid isPermaLink="true">https://windtask.com/articles/playwright-chromium-lag-coding-agents/</guid><description>Chrome ships fixes; Playwright pins Chromium later. Decide by workload—own-app E2E, agent untrusted browsing, Playwright MCP—using chromiumSandbox, channel chrome, non-root hosts, and egress allowlists. A decision matrix for Claude Code, Cursor, and MCP.</description><pubDate>Wed, 04 Nov 2026 00:00:00 GMT</pubDate><category>Security</category><category>playwright</category><category>chromium</category><category>coding-agents</category><category>claude-code</category><category>cursor</category><category>playwright-mcp</category><category>browser-security</category><category>chromium-sandbox</category></item><item><title>pull_request_target Default Block: A Dec 2 Migration Runbook</title><link>https://windtask.com/articles/github-pull-request-target-default-block/</link><guid isPermaLink="true">https://windtask.com/articles/github-pull-request-target-default-block/</guid><description>Public GitHub repos block pull_request_target by default on December 2, 2026, with brownouts on November 9 and 16. Inventory workflows, switch CI to pull_request, and allow the trigger only for jobs that never execute fork code.</description><pubDate>Wed, 04 Nov 2026 00:00:00 GMT</pubDate><category>Security</category><category>github-actions</category><category>pull-request-target</category><category>ci-security</category><category>workflow-policy</category><category>secrets</category></item><item><title>Shared Firefox Sessions for Coding Agents: When Kitsune-Style MCP Is Worth the Blast Radius</title><link>https://windtask.com/articles/kitsune-firefox-mcp-coding-agents/</link><guid isPermaLink="true">https://windtask.com/articles/kitsune-firefox-mcp-coding-agents/</guid><description>Decide when a coding agent should share your real Firefox logins versus driving a separate automation browser. Compare Kitsune, firefox-bridge, Playwright MCP, and Claude in Chrome on session model, profiles, and approval gates—without pretending modes stop prompt injection.</description><pubDate>Wed, 04 Nov 2026 00:00:00 GMT</pubDate><category>Security</category><category>kitsune</category><category>firefox-mcp</category><category>coding-agents</category><category>claude-code</category><category>playwright-mcp</category><category>browser-security</category><category>shared-session</category></item><item><title>Coding-Agent Cloud Waste: How to Find and Contain What Your Agents Leave Running</title><link>https://windtask.com/articles/coding-agent-cloud-waste-agent-leftovers/</link><guid isPermaLink="true">https://windtask.com/articles/coding-agent-cloud-waste-agent-leftovers/</guid><description>Coding agents create VMs, databases, and deploys mid-task—then sessions end without teardown. A practical FinOps playbook for read-only audit, human-confirmed cleanup, Claude Code/Codex/MCP blind spots, and what --verify can (and cannot) prove.</description><pubDate>Tue, 03 Nov 2026 00:00:00 GMT</pubDate><category>FinOps</category><category>coding-agents</category><category>cloud-waste</category><category>claude-code</category><category>codex</category><category>mcp</category><category>finops</category></item><item><title>Sample: categories are paths, not subdomains</title><link>https://windtask.com/articles/sample-categories-are-paths/</link><guid isPermaLink="true">https://windtask.com/articles/sample-categories-are-paths/</guid><description>A stub so the site has a second category. The FinOps essay is the real article on windtask.com.</description><pubDate>Tue, 03 Nov 2026 00:00:00 GMT</pubDate><category>Notes</category><category>meta</category></item></channel></rss>