RepoRadar
Leave a Tip
Item detail

corsairdev/corsair

Corsair corsairdev/corsair is an Apache-2.0 open-source unified integration layer for AI agents that exposes a single MCP (Model Context Protocol) endpoint with scoped per-tool permissions (open, cautious, strict, readonly) over Gmail, Slack, GitHub, Google Drive, Notion, and other SaaS, intercepting every destructive action behind a human-approval link that the operator must click before the agen

Open source →Back to radar
Score7.4
Popularity2781.0
Riskmedium
TierSilver
Score breakdown
Usefulness7.4
Novelty10.0
Momentum10.0
Maturity7.9
Open-source/build7.4
Evidence7.2
Workflow potential7.4
Setup ease6.5

Popularity is tracked separately. Support, ads, sponsorships, and tips never affect these signals.

Why it matters

Useful for AI engineers and agent builders who want to wire a Claude/Codex/GPT-4-class agent into Gmail/Slack/GitHub/Notion/Drive without giving it raw OAuth tokens or letting destructive calls run unmonitored, because Corsair corsairdev/corsair ships as a single MCP server with per-integration permission modes (open / cautious / strict / readonly), human-approval links for destructive actions, an

Who should use it

BuildersPower users

Who should skip it

Skip or sandbox it if you cannot review permissions, data access, and failure modes before use.

Risk explanation

Medium risk: use sandboxing, least privilege, and explicit review before connecting sensitive data or accounts.

Evidence links

Closest alternatives / related signals

  • No close related item is available in the current public catalog.