hn.today

Show HN: Pulsar – cross-harness agent to agent communication and messaging

github.com20 points4 comments
Screenshot of Show HN: Pulsar – cross-harness agent to agent communication and messaging

Pulsar is an MCP server for communication among AI coding agents running in different harnesses on the same machine, including Claude Code, Codex, OpenCode, pi, omp, Cursor and Gemini CLI. It maintains a shared view of active agents, with automatically detected details such as harness, project, branch, worktree and pull request, plus agent-defined tags and activity descriptions. Agents can find recipients by name or selectors, send messages to individuals or groups, broadcast to everyone, and use threaded replies, blocking questions and waits for messages or changes in another agent’s status. Messages can wake idle agents, and urgent delivery can reach an agent mid-turn where its harness supports that.

A lightweight daemon, started on demand, keeps the agent registry and a SQLite message log; harness-specific adapters push messages into sessions, while unsupported clients receive them in an inbox. The setup command detects installed harnesses, previews and backs up files it will change, and installs the server locally; it also supports dry runs, selective installation and undo. Pulsar’s central advantage over Claude-only messaging is cross-harness discovery, group addressing and searchable receipts and history, though its first version is limited to one machine. Its safety model treats local processes running as the user as trusted: a user-only socket defines access, sender identities are assigned by the daemon, and messages are marked as peer input rather than user instructions. Agents can accept, hold or refuse messages, which cannot grant permissions; rate limits apply and telemetry is not collected.

Read on github.com4 comments on Hacker News

Summary generated by AI from the linked article. hn.today is not affiliated with Hacker News or Y Combinator.

More in AI

Human.md

Human.md

Human.md argues that people working alongside increasingly capable AI should bring intentional rule-breaking, unexpected inspiration, empathy, emotional resonance, and delight. It urges creators to experiment and ship many prototypes rather than rely on polished opinions or merely meet a brief. (human-md.ai)

The daily digest

Today's best Hacker News stories, summarized and screenshotted, one email a day.