v1.3.1

Maintenance — the recommended Anthropic model is current again.

Changed

• recommended_models now names anthropic:claude-opus-5 (was anthropic:claude-opus-4-8, a generation behind — the Claude 5 family is current). Updated in the persona frontmatter and in the "which models work best?" FAQ of the help guide ×17. openai:gpt-5.5 is unchanged.

This is a recommendation only: OpenWorker still runs whatever model you have configured, including a local Ollama one, so nothing changes for an existing install. Re-import the persona to pick it up.

Verified

• scripts/validate-persona.py — INSTALLABLE ✓ (OpenWorker manifest rules).
• Help-bundle parity ×17 — 12 H2 headings, structure unchanged.
• Every technical claim in the persona re-checked line by line against the live projects: all six npm run scripts (scan, tracker, find, patterns, verify:portals, cv:verify-facts) and their .mjs targets exist in Fighter90/career-ops; every scan.mjs flag the persona documents (--dry-run, --company, --since, --posted-after, --verify, --quiet) is real; web-ui/bin/start.sh and port 4317 are current; config/two-pager.yml, config/memory.md, data/applications.md and modes/ all exist. No other drift found.

career-ops-coworker.zip attached for OpenWorker → Install a coworker → .zip.