Player profile
Alfred
GitHub issues, in. Pull requests, out. A self-hosted runtime for autonomous Claude Code and Codex agents on the CLI subscriptions you already pay for.
Profile
Where it fits
Alfred is an MIT-licensed Python runtime for autonomous engineering agents. Each agent is a narrow role (planner, implementer, reviewer, tester) backed by your own Claude Code or Codex CLI subscription. No provider API keys, no cloud agent service, no second LLM bill.
Work intake is GitHub-native: scoped issues and specs define what to do, labels (agent:implement, agent:in-flight, agent:pr-open, agent:done) hold state, and each firing runs in a fresh git worktree. Results land as reviewed pull requests, follow-up tests, and Slack summaries.
Alfred is the operating layer around the CLI coding agents you already use: scoped intake, role-based engine routing, worktree isolation, bounded autonomy, review handoff, and scheduled runs across one repo or many.
Builder web analytics
Measure projects built with Alfred
Alfred can ship product changes, docs, tests, and review fixes through scheduled agent runs. Agent Analytics gives a follow-up agent traffic, source, funnel, and conversion data to judge whether the shipped change helped users.
Install Agent Analytics on the app, docs, or marketing surface affected by Alfred-created PRs. Alfred's GitHub labels and shipped summaries provide the work log; Agent Analytics supplies user behavior after deploy.
First loop to measure
- Alfred claims a scoped GitHub issue and runs a Claude Code or Codex job in an isolated worktree
- the resulting PR ships a product change, docs update, test coverage, or review fix
- the deployed surface reports visits, sources, signup, activation, retention, funnels, and conversion events to Agent Analytics
- a follow-up agent compares the changed path with the prior period and reports what improved or regressed
Copyable prompt
Use Agent Analytics for this project. If event reporting is missing, add the tracker and report events for this project surface, including changes shipped by Alfred agents. Verify events are arriving. Then fetch the last 7 days and compare them with the prior 7 days. Tell me which changes shipped by Alfred agents moved users toward value, where users dropped off, which sources mattered, and what my agent workflow should improve next.Related