Follow Discord

DisclaimerUnofficial, and not affiliated with Anthropic. Nearly all of this is read straight out of what ships: npm bundles, captured prompts, published docs. Anthropic's own notes go in verbatim, marked as theirs. The rest is my reading, and every entry carries the strings behind it. If one looks wrong, vote it down and say why.

← All of v2.1.81 Home All releases ←olderv2.1.80 v2.1.83newer→

Claude Code v2.1.81 20 Mar 2026

MCP Tool Read/Search Allowlist [In Development]

What

A comprehensive allowlist classifying hundreds of known MCP tool names as "search" or "read" operations, likely for auto-approving read-only MCP operations.

Status

Infrastructure added, not yet visibly wired to auto-approval logic.

Details

Search tools: ~90 tools from Slack, GitHub, Datadog, Sentry, Google, Atlassian, Stripe, PubMed, Brave, and more
Read tools: ~270 tools covering Slack reads, GitHub detail views, Grafana dashboards, PagerDuty, Jira/Confluence, Asana, MongoDB, Elasticsearch, Figma, Puppeteer screenshots, and more
Classification uses snake_case normalization of tool names

Evidence

MCP tool classification allowlist — PT4() at line ~327832 (search for "slack_search_public")

Strings lifted out of the shipped bundle, so the claim above can be checked against them.

See this entry in the whole of v2.1.81 →