Claude Code Daily Briefing - 2026-03-22

Release Summary

VersionDateKey Changes
v2.1.813/20--bare scripted execution mode, --channels mobile permission relay, OAuth concurrent session fix, voice mode stabilization (latest)
v2.1.803/19--channels MCP message push (research preview), rate_limits statusline, inline plugin declarations, ~80MB memory savings on large repos
v2.1.793/18--console auth flag, turn duration toggle, VSCode /remote-control, ~18MB memory savings

Full Release Notes


New Features & Practical Usage

DOJ Directly Counters Anthropic’s First Amendment Argument — 40-Page Filing (3/17, Key This Week)

The Department of Justice filed a 40-page brief urging the California federal court to reject Anthropic’s preliminary injunction request. Legal battle reaches peak intensity ahead of the 3/24 hearing.

DOJ’s core arguments:

The Hill | TechStory

Pentagon CTO “Pretty Confident” About Life After Anthropic (3/18)

Pentagon Under Secretary (CTO) Emil Michael publicly expressed confidence in replacing Anthropic.

Key quotes:

Breaking Defense | Federal News Network

Anthropic Fires Back with “Nearly Aligned” Email and Sabotage Denial (3/20–21)

Anthropic countered with two sworn declarations contradicting Pentagon’s “national security threat” claims, plus an official sabotage denial.

Latest developments:

TechCrunch | National Today


Developer Workflow Tips

Claude Code Channels — Control from Telegram in 5 Minutes

Channels, introduced as a research preview in v2.1.80, let you send messages from Telegram or Discord directly into a running Claude Code session.

# Start Claude Code with channel support
claude --channels

Built on MCP, so the community can build additional connectors for Slack, WhatsApp, etc. v2.1.81’s --channels permission relay enables mobile tool approval.

--bare Mode — The Key to CI/CD Automation

ANTHROPIC_API_KEY=sk-... claude -p "Review this PR" --bare

Skips hooks, LSP, plugin sync, and skill directory walks for fast, clean execution in automated environments.


Security & Limitations

With the preliminary injunction hearing two days away, both sides have fully revealed their legal arsenals.

Position comparison:

IssueAnthropicDOJ/Pentagon
First AmendmentAI safety policy is protected speechContract term refusal is conduct, not speech
National security threat”Nearly aligned” email proves contradictionTerms of service became unacceptable
Sabotage concernsNo backdoor/kill switch, sworn under oathSupply chain “pollution” risk
ReplaceabilityHundreds of millions in revenue at risk”Already deployed OpenAI, confident in transition”
Industry supportMicrosoft, 150 retired judges, competitor staffSovereign national security judgment

3/24 key question: Judge Rita Lin will decide whether to pause the supply-chain risk designation.

3/21 Claude Opus & Sonnet 4.6 Service Incident — Continued Monitoring

Yesterday’s (3/21) elevated error rates affecting Opus/Sonnet 4.6 have been resolved. Stability monitoring continues today.

Claude Status


Ecosystem & Plugins

Claude Code Channels Community Adoption Accelerating (3/20–22)

After hitting #4 on HN with 341 points, community adoption is ramping up rapidly.

VentureBeat | DEV Community

Claude Double Usage Promotion — 6 Days Left (3/13–3/28)

Anthropic’s usage doubling promotion runs through March 28.

Claude Help Center


Community News


Minor Changes



Tomorrow’s Watch List