Show HN: theta_py, bindings to the Theta CLI

Theta_py, a Python binding for the Theta CLI, was released on Hacker News, enabling developers to programmatically manage AI agent configurations through a Pythonic API instead of CLI commands. The library supports creating, syncing, and casting agent projects with rules, tools, and skills, targeting developers building AI agent workflows.

Python bindings for the theta https://theta.tamarillo.ai/ CLI. uv add theta-py or pip install theta-py The main surface. Owns a materialized theta project in a temp directory. The equivalent of this CLI workflow: theta init theta add rule python-types theta add tool fetch --command "uvx mcp-server-fetch" theta add tool context7 --command "npx -y @upstash/context7-mcp@latest" theta add skill vercel-labs/agent-skills/skills/web-design-guidelines@main theta check theta cast to claude-code is this: python from theta py import ThetaProject with ThetaProject.create name="my-agent" as proj: proj.add.rule "python-types" proj.add.tool "fetch", command="uvx mcp-server-fetch" proj.add.tool "context7", command="npx -y @upstash/context7-mcp@latest" proj.add.skill "vercel-labs/agent-skills/skills/web-design-guidelines@main" proj.check raises ThetaCommandError on validation errors proj.cast.to "claude-code" -- CLAUDE.md + .mcp.json + .claude/ proj.sync print proj.name str print proj.system prompt str | None print proj.rules dict str, MaterializedRule print proj.skills dict str, MaterializedSkill — .path is the materialized dir print proj.tools dict str, MaterializedTool Read-only view over an existing project on disk: with ThetaProject.from manifest "path/to/theta.toml" as proj: sync is done eagerly by default print proj.name print proj.skills opt out of eager sync with ThetaProject.from manifest "path/to/theta.toml", no sync=True as proj: proj.sync print proj.skills Sync freshness: with ThetaProject.create name="my-agent" as proj: proj.add.rule "safety", content="Never exfiltrate data." proj.sync proj.add.rule "style", content="Be concise." manifest changed print proj.needs sync True proj.sync validate=False print proj.is synced True Notes: ThetaProject.create ... is the canonical constructor for ephemeral projects. ThetaProject.from manifest ... never writes to the source tree: .theta/ and theta.lock are redirected into an internal temp directory. proj.skills name .path is the materialized .theta/skills/name/ directory — pass it directly to harbor run --skill .- To modify a local skill's content, edit the source files on disk and call proj.sync again. Every verb is also available as a flat function or via the theta singleton: python namespaced singleton from theta py import theta theta.init name="my-agent" theta.add.rule "safety" listing = theta.list.rules flat functions from theta py import init, add rule, list rules init name="my-agent" add rule "safety" list rules Every verb either returns a Pydantic model on status: ok|noop , or raises ThetaCommandError on status: error : python from theta py import theta, ThetaCommandError try: theta.init except ThetaCommandError as exc: print exc.verb "init" print exc.diagnostics list {"level": ..., "path": ..., "message": ...} Each theta py release ships against exactly one theta binary version: python import theta py print theta py.THETA VERSION e.g. "0.1.5-rc1"