upstream: official claude code updates 2026-06-16

[github-actions]

upstream changes detected

sources: anthropic/changelog, anthropic/official-docs, anthropics/claude-code, marckrenn/claude-code-changelog, shanraisshan/claude-code-best-practice
changes found: 14
files edited: 5
staleness flags: 5

what changed upstream

• [release] v2.1.179 (link)
• [commit] Merge pull request feat(cc)!: v3.5.0 — observability (CC_DEBUG, CC_TRACE_SQL) + digest verbosity by $CLAUDE_EFFORT #117 from shanraisshan/settings-drift/2026-06-16 (link)
• [commit] v2.1.179 (link)
• [commit] v2.1.179 (link)
• [commit] v2.1.179 - init (3,406 tokens) (link)
• [commit] v2.1.179 - init (1,774 tokens) (link)
• [commit] v2.1.179 - init (1,524 tokens) (link)
• [commit] v2.1.179 - edit (link)
• [commit] v2.1.179 - edit (link)
• [commit] v2.1.179 - edit (link)
• [commit] v2.1.179 - edit (link)
• [commit] v2.1.179 - edit (link)
• [changelog] Claude Code changelog updated (link)
• [docs_change] Official Claude Code docs updated (link)

staleness flags

file	claim	reality	severity
docs/tips/fast-mode.md	fast mode now coexists with effort levels (--effort low	medium	high
docs/session-workflow.md	the welcome screen stacking multiple promotional banners — at most one promo now shows per session	v2.1.179 fixed multiple promotional banners stacking; docs don't mention this UI behavior at all	low
docs/hooks.md	hook fire frequency is driven by tool usage. from real session data: Bash (10,153) is most hook-triggering	v2.1.179 added Tool(param:value) syntax for permission rules (e.g. Agent(model:opus)) which affects PreToolUse hook behavior. docs show old pattern.	medium
docs/tips/hooks-v2.md	hook types: command, http, prompt, agent, mcp_tool (v2.1.118+)	v2.1.179 docs mention Tool(param:value) syntax for matching tool input parameters in permission rules. hooks-v2.md doesn't document this matching capability.	medium
docs/agents.md	worktree isolation and agent spawning patterns described for v2.1.122	v2.1.178+ added nested .claude/ directories support and skill nesting. agents.md doesn't mention .claude/skills/ nesting or directory-scoped skill resolution.	medium

proposed edits

• docs/hooks.md: v2.1.178 introduced Tool(param:value) matching for permissions; hooks.md documents safety but not the new matching capability
• docs/agents.md: v2.1.178 added automatic nested .claude/skills/ loading; agents.md shows agent patterns but doesn't mention skill nesting
• docs/tips/context-management.md: v2.1.178 improved nested .claude/ directory resolution; context-management.md covers v2.1.160 but not v2.1.178 workflow improvements
• docs/session-workflow.md: v2.1.179 fixed promotional banner stacking; session-workflow.md covers session end but not UI changes
• docs/cost.md: v2.1.179 fixed mid-stream response preservation; cost.md covers monitor reliability but this fix is worth noting

review checklist

• staleness flags accurate?
• version numbers and pricing correct?
• proposed edits read naturally?

auto-generated by official-watcher