nanocoai/nanoclaw / .claude/skills/add-codex
add-codex
FlaggedUse Codex (OpenAI's codex app-server) as a full agent provider — planning, tool orchestration, MCP tools, server-side history, session resume — alongside or instead of Claude. ChatGPT subscription or OpenAI API key, vault-only via the selected gateway. Per-group via `ncl groups config update --provider codex`. Distinct from using OpenAI as an MCP tool (where Claude remains the planner).
payload/src/providers/codex-agents-md.ts
View on GitHub/**
* AGENTS.md spec for codex agent groups — the MIXED-VERSION COMPATIBILITY path.
*
* On a core with the provider contracts, AGENTS.md instructions come from the
* core-owned canonical template rendered with the instruction facts the codex
* host contract declares (`src/provider-contracts/codex.ts`); this module is
* never reached there. On a core that predates the contracts, the legacy host
* adapter composes AGENTS.md through this spec, whose extra sections and base
* document (`container/AGENTS.md`) preserve the exact document those cores
* always produced. Both core generations still type `baseDocPath` and
* `extraSections` on `ProjectDocSpec` (deprecated on the contract core), so
* the spec is typed directly.
*
* `project_doc_max_bytes` is mirrored in the container provider's config.toml
* writer. Over the cap the shared composer degrades — it drops the largest
* capability sections, logs what went, and says so in the document — rather
* than throwing, which would ride `wakeContainer`'s retry contract and dark
* the group.
*/
import path from 'path';
import { composeGroupProjectDoc, type ProjectDocSpec } from '../project-doc-compose.js';
import type { AgentGroup } from '../types.js';
// Codex's own default is 32 KiB, but a default Slack install composes about
// 39 KB, and over the cap the composer drops whole sections (#4464).
export const CODEX_PROJECT_DOC_MAX_BYTES = 64 * 1024;
export const CODEX_PROJECT_DOC_EXTRA_SECTIONS = [
{
name: 'Memory System',
body: [
'The live memory index and definition are supplied by NanoClaw at session startup, clear, and after compaction.',
'Editable memory-system definition: `/workspace/agent/memory/system/definition.md`.',
'Top memory index: `/workspace/agent/memory/index.md`.',
'Read the definition and index, then use linked memory files and conversation archives when relevant.',
'Stored user preferences are binding: read any linked memory file relevant to the user or the request, and apply it without being asked.',
'Do not use `AGENTS.local.md` or `AGENTS.override.md` for memory.',
].join('\n\n'),
},
{
name: 'Native Runtime Skills',
body: [
'Selected NanoClaw runtime skills are available as Codex-native skills at `/workspace/agent/.agents/skills`.',
'Each skill directory contains a `SKILL.md` with its trigger description plus any supporting files, and points to the read-only shared skill source under `/app/skills`.',
'Use skill discovery to load these skills only when their descriptions match the task. A skill whose rules must hold before the task is recognised ships an `instructions.md` instead, and those arrive inlined as `NanoClaw Skill:` sections of this document.',
'Skills YOU author or install yourself go in `~/.codex/skills/<name>/SKILL.md` — persistent across sessions and discovered by Codex automatically. Never write skills elsewhere: paths outside `~/.codex` and `~/.agents` are ephemeral or not discovered.',
].join('\n\n'),
},
];
const CODEX_PROJECT_DOC: ProjectDocSpec = {
fileName: 'AGENTS.md',
baseDocPath: path.join('container', 'AGENTS.md'),
extraSections: CODEX_PROJECT_DOC_EXTRA_SECTIONS,
maxBytes: CODEX_PROJECT_DOC_MAX_BYTES,
};
export function composeGroupAgentsMd(group: AgentGroup, groupDir: string): Promise<void> {
// Pre-contract cores read baseDocPath/extraSections; contract cores ignore
// them (and never call this path for surfaces).
return composeGroupProjectDoc(group, groupDir, CODEX_PROJECT_DOC);
}
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| reverse-engineer-anythingmorluto/reaReverse engineer native, managed, Electron/JavaScript, packaged, firmware, and browser targets with REA, and build agent environments (AgentEnv bundles) of web applications from a live instance or a capture, and of Linux desktop apps with a task from a request. Use shipped-artifact or requested runtime evidence to explain features, compare versions, decompile code, or guide a reconstruction. Skip REA for ordinary source-repository architecture analysis. | Stars: 471k | Updated | Claude CodeCodex | Verified | Install |
| diagnosing-superpowersobra/superpowersUse when a superpowers session went wrong and your human partner wants to know why — repeated work, ignored plans, stumbles, poor results, a skill that didn't fire, "it took too long", "why is it so expensive", "what is it doing" — or wants to build a bug report for the superpowers maintainers, for the current session or a past one identified by id or path, on any harness. | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |
| dispatching-parallel-agentsobra/superpowersUse when facing 2+ independent tasks that can be worked on without shared state or sequential dependencies | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |
| executing-plansobra/superpowersUse when executing an implementation plan in the current session as the implementer yourself — your human partner chose inline execution, or no subagent tool is available | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |