agentsecrets and agent-secrets
These are competitors offering alternative approaches to AI agent credential management—The-17/agentsecrets emphasizes zero-knowledge infrastructure for agent operations while joelhooks/agent-secrets focuses on portable credential management with encryption and lifecycle controls, but both solve the same core problem of securing secrets for AI agents.
About agentsecrets
The-17/agentsecrets
Zero-knowledge secrets infrastructure built for AI agents to operate, not just consume.
Supports transport-layer credential injection via HTTP proxy, MCP server, and environment variable wrapping—keeping credential values structurally outside agent context. Built around X25519 key exchange and AES-256-GCM encryption with client-side key derivation, ensuring the server never holds decryptable secrets. Integrates with Claude Desktop, Cursor, and any HTTP-capable agent framework, with governance audit logging, team workspaces, and an SDK for building credential-aware tools.
About agent-secrets
joelhooks/agent-secrets
🛡️ Portable credential management for AI agents — Age encryption, session leases, killswitch
Implements a Unix socket daemon architecture with JSON-RPC for CLI communication, storing secrets encrypted at rest with age and issuing time-bounded leases tied to client identifiers for granular audit trails. Built in Go with append-only logging and pluggable rotation hooks (e.g., `gh auth refresh`), it integrates with AI agent frameworks via the OpenClaw skill system and exposes self-documenting JSON command trees for agent discovery.
Scores updated daily from GitHub, PyPI, and npm data. How scores work