Who contributed to Anthropic on May 21, 2026?

4 developers shipped this update, including claude[bot], iainmcgin, mattpic-ant, and dlaw-ant.

What were the notable Anthropic updates?

Bump mercadopago to 63ff263c (v2 + PreToolUse hook gating) (#1949), Bump mercadopago to 63ff263c (v2 + PreToolUse hook gating), and codegen: address style audit findings — clones, wildcards, unwrap (#121).

@anthropics

Claude SDKs and developer tooling

github ↗

Pick a date

Topics: AI / ML Full archive →

The Wire · Showcase

MERCADOPAGO V2 GOES LIVE, BUFFA AUDIT CLEARS CODEGEN DEBT

By RepoJournal · Filed 06:02 UTC on May 21, 2026 · About Anthropic

4 people shipped this

iainmcgin @iainmcgin 2 cited

claude[bot] @claude[bot] 1 cited

mattpic-ant @mattpic-ant 1 cited

dlaw-ant @dlaw-ant 1 cited

MercadoPago plugin ships consolidated v2 architecture with HTTP MCP endpoints and smarter hook gating, while Buffa's codegen passes style audit with redundant clone removal across 6 sites.

The mercadopago plugin pins to 63ff263c, completing the v2 migration that replaces the keychain-read wrapper with a plain HTTP MCP entry and collapses 13 skills into 4 orchestration skills [1] [2]. That same SHA includes the May 19 PreToolUse hook fix that stops the hook from running on unrelated projects, cutting noise in agent bundles. Over in Buffa, the codegen audit wrapped up with two follow-up PRs clearing unnecessary Option/Result wraps and redundant clones [3] [4] [5]. The team also shipped linked descriptor types and dedup feature resolution in the descriptor module [6], laying groundwork for runtime reflection in subsequent PRs. Meanwhile, financial-services fixed 9 failing marketplace loader syncs by synthesizing proper YAML frontmatter across 29 vertical-plugin SKILL.md files [7] [8], and stamped metadata.anthropic_cookbook on managed-agent deploys for BigQuery attribution [9] [10]. Claude Agent SDK rolled to v0.2.83 with CLI 2.1.146 bundled in [11] [12] [13], and Claude Code Action hit v1.0.128 with matching CLI and SDK bumps [14] [15].

Action items

→ Sync mercadopago plugin to 63ff263c if using PreToolUse hooks in multi-project contexts anthropics/claude-plugins-official [plan]
→ Verify SKILL.md frontmatter in vertical plugins if deploying to marketplace anthropics/financial-services [plan]
→ Update Claude Agent SDK to 0.2.83 for bundled CLI alignment anthropics/claude-agent-sdk-python [monitor]

References

[1] Bump mercadopago to 63ff263c (v2 + PreToolUse hook gating) (#1949) anthropics/claude-plugins-official
[2] Bump mercadopago to 63ff263c (v2 + PreToolUse hook gating) ↗ anthropics/claude-plugins-official
[3] codegen: address style audit findings — clones, wildcards, unwrap (#121) anthropics/buffa
[4] codegen, text: remove vestigial Option/Result wraps and redundant clones (#123) anthropics/buffa
[5] codegen: address style audit findings — clones, wildcards, unwrap ↗ anthropics/buffa
[6] descriptor: add linked descriptor types and dedup feature resolution ↗ anthropics/buffa
[7] fix(plugins): add SKILL.md frontmatter, repair hooks.json, dedupe spglobal MCP ↗ anthropics/financial-services
[8] fix(plugins): add SKILL.md frontmatter, repair hooks.json, dedupe spglobal MCP (#243) anthropics/financial-services
[9] Stamp metadata.anthropic_cookbook on managed-agent deploys ↗ anthropics/financial-services
[10] Stamp metadata.anthropic_cookbook on managed-agent deploys (#245) anthropics/financial-services
[11] v0.2.83 ↗ anthropics/claude-agent-sdk-python
[12] docs: update changelog for v0.2.83 anthropics/claude-agent-sdk-python
[13] chore: bump bundled CLI version to 2.1.146 anthropics/claude-agent-sdk-python
[14] v1.0.128 ↗ anthropics/claude-code-action
[15] chore: bump Claude Code to 2.1.146 and Agent SDK to 0.3.146 anthropics/claude-code-action

Quick answers

What shipped in Anthropic on May 21, 2026?: MercadoPago plugin ships consolidated v2 architecture with HTTP MCP endpoints and smarter hook gating, while Buffa's codegen passes style audit with redundant clone removal across 6 sites. In total, 16 commits, 13 pull requests, and 2 releases landed.
Who contributed to Anthropic on May 21, 2026?: 4 developers shipped this update, including claude[bot], iainmcgin, mattpic-ant, and dlaw-ant.
What were the notable Anthropic updates?: Bump mercadopago to 63ff263c (v2 + PreToolUse hook gating) (#1949), Bump mercadopago to 63ff263c (v2 + PreToolUse hook gating), and codegen: address style audit findings — clones, wildcards, unwrap (#121).

CLAUDE AGENT SDK FIXES STDIN CLOSURE BUG IN BACKGROUND TASKS

The Python SDK's query method was prematurely closing stdin when background subagents were still in flight, breaking tool calls and silently bypassing hooks - now fixed in v0.2.127.

ai 16 shipped 1-min read

@anthropics 1 day ago

BUFFA v0.9.1 PATCHES REGRESSION THAT BROKE ORDINARY SCHEMAS

Buffa shipped an emergency patch overnight after v0.9.0 broke codegen on schemas with a few hundred proto files, small enough to hit in production.

ai 113 shipped 1-min read

@anthropics 2 days ago

CLAUDE FOUNDATION MODELS ADDS APP ATTEST, PLUGIN ECOSYSTEM STABILIZES

iOS developers can now ship Claude calls without API keys in the binary, backed by Apple's Secure Enclave.

ai 111 shipped 1-min read

@anthropics 3 days ago

BUFFA FIXES V0.9.0 REGRESSION, CLAUDE PLUGINS ECOSYSTEM GAINS 9 NEW INTEGRATIONS

Buffa's codegen was silently rejecting moderately large schemas due to a memory bound meant only for untrusted input, but a fix shipped overnight restores throughput and clears the path for googleapis-scale proto sets.

ai 156 shipped 2-min read

Elsewhere on the wire

AI Agents about 10 hours ago

CLAUDE OPUS 5 LANDS ACROSS THE STACK

The newest Anthropic model is now live in langchain, Cline, and llama-index, with native support for extended reasoning and 1M context windows.

ai-agents 28 shipped 1-min read

Local LLMs about 10 hours ago

OLLAMA LANDS LAGUNA SUPPORT AND CRUSHES MEMORY LEAKS WHILE SGLANG HITS V0.5.16 WITH CONFIDENCE-DRIVEN SPECULATIVE DECODING

Ollama shipped three critical performance and reliability fixes for Metal residency and concurrent access patterns, while SGL-Lang released 0.5.16 with a new speculative algorithm hitting 383.7 tok/s on DeepSeek-V4.