v1.1.x Changelog

This version series introduces the local credential proxy, response body redaction, command-line environment injection, and cryptographic Agent Identities.

v1.1.3

• Workspace Config Fallbacks: Added fallback checks to resolve global project configurations in target commands when local workspace settings are omitted.
• OpenClaw Exec Provider: Overhauled the OpenClaw integration tool wrapper to execute via the native agentsecrets binary rather than fallback Python scripts.

v1.1.2

• Multi-Environment Scoping: Promoted development, staging, and production to first-class concepts in the CLI. Switch contexts instantly via agentsecrets environment switch.
• Keychain-Auth Agent: Integrated local keychain-auth sockets to cache user authorization keys in secure OS memory contexts.

v1.1.1

• Agent Identity Management: Added commands to declare agent names and request cryptographic verification tokens (agent list, agent token issue).
• Proxy Sync and Streaming: Integrated real-time proxy event logging, streaming summary breakdowns, and PID-level monitoring.

v1.1.0

• Environment Variable Injection: Introduced the agentsecrets env -- <command> command to inject credentials straight into running subprocess memory (e.g. Node.js or Python processes) without writing to disk.
• Zero-Trust Credential Proxy: Added a local HTTP proxy server supporting bearer tokens, basic auth, headers, forms, and JSON body injection.
• Response Body Redaction: The local proxy now intercepts API responses and redacts matching credential values, replacing them with [REDACTED_BY_AGENTSECRETS] to keep secrets out of AI agent logs.
• Domain Allowlist: Restricts the proxy from forwarding calls to unauthorized third-party URLs. Modifying allowlists requires admin privileges and local password confirmation.