Skill Claude CodeCodex
Use a project's FastFollowSpec to produce source-grounded research, target gap assessments, issue candidates, or separately admitted implementations without treating upstream projects as authority.
Skill Claude CodeCodex
Join a public NIP-29 Nostr chat with your own key, the nak command-line tool, and standard relay frames.
Instructions file CodexOpenCode
Instructions for OpenAgentsInc/openagents, covering openagents agent contract, scope, simplified technical english, writing and proactive subagent delegation (owner mandate).
Instructions file
Instructions for OpenAgentsInc/openagents, a project described as: Monorepo & docs.
Agent
Date: 2026-06-16 Status: architecture audit, no implementation yet Scope: Flue as a possible agent harness and third-party connector layer for OpenAgents product, private workspaces, operator workflows, and launch/business operations.
Agent
Date: 2026-06-16 Status: implementation roadmap, no implementation yet Scope: GitHub-first Flue connector integration tied into the apps/openagents.com Worker and web UI.
Agent
Date: 2026-06-17 Status: architecture audit, no implementation yet Scope: the local Cloudflare Agents SDK clone at projects/repos/agents, plus current Cloudflare Agents documentation, as a possible OpenAgents runtime for stateful agents, workrooms, connector events, realtime UI, and durable turns.
Agent
Date: 2026-06-17 Status: architecture audit, no implementation yet Scope: Rivet's June 16, 2026 Effect SDK announcement and the local projects/repos/rivet clone as a possible OpenAgents actor/runtime substrate.
Skill Claude CodeCodex
Use the OpenAgents CLI to read and write issues, projects, repositories, and the forum, to call any API route, to sign a person in, and to run Autopilot. Use it whenever the work touches OpenAgents itself rather than the files in this repository, or when the person says run Autopilot.
Skill Claude CodeCodex
How to hand work to another agent, run Autopilot, and work through an issue backlog with safe parallelism. Loaded into every session.
Skill Claude CodeCodex
Canonical structured-report invocations for suite runners, so failure names are read from a file instead of recovered by re-running a suite. Use whenever a test run might fail, or whenever you need the list of failing tests from a run you already paid for.