add-codex

Use 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 OneCLI. Per-group via `ncl groups config update --provider codex`. Distinct from using OpenAI as an MCP tool (where Claude remains the planner).

Quality

72%

Does it follow best practices?

Run evals on this skill

Adds up to 20 points to the overall score

View guide

Securityby

Passed

No findings from the security scan

Fix and improve this skill with Tessl

tessl review fix ./.claude/skills/add-codex/SKILL.md

Quality

Content

77%Weight 40%Scale 1-3

Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.

The body is highly actionable with a clear, validated multi-step workflow and concrete executable directives throughout, scoring top marks on actionability and workflow clarity. It loses points on conciseness for inline version pins and some explanatory padding, and on progressive disclosure because all detail lives inline with no in-bundle references to offload it.

Suggestions

Move the pinned Codex CLI version (and any other time-sensitive pins) into a clearly marked version/config section or a referenced file so the main flow stays evergreen and conciseness is not penalized.

Consider extracting the long Pre-flight checklist and the copy payload file list into a referenced file (e.g., references/payload-files.md) to add genuine one-level-deep progressive disclosure.

Tighten the prose explaining the nc: directive system and the AGENTS.md compose mechanics to remove explanation an agent likely infers from context.

Dimension	Reasoning	Score
Conciseness	The body is mostly lean and assumes competence (no tutorial padding on barrels/JSON-RPC), but inline time-sensitive material — notably the pinned "0.138.0" version declared as canonical — and some explanatory prose around the nc: directive system could be tightened, which the guideline says should penalize conciseness.	2 / 3
Actionability	Every step ships concrete executable directives (nc:copy/append/json-merge/run) with exact file lists, exact shell commands, and a pinned version, making the guidance copy-paste ready rather than descriptive.	3 / 3
Workflow Clarity	A clear Pre-flight skip-check feeds into numbered steps 1-5 ending in an explicit Validate checkpoint, with registration tests that go red on drift and a Troubleshooting section giving error-recovery feedback loops.	3 / 3
Progressive Disclosure	No bundle files exist and sections are well organized, but the document is self-contained with no one-level-deep in-bundle references; sizable material such as the copy payload list and CLI manifest entry is inline rather than split into referenced files.	2 / 3
	Total	10 / 12 Passed

Description

67%Weight 40%Scale 1-3

Based on the skill's description, can an agent find and select it at the right time? Clear, specific descriptions lead to better discovery.

The description is specific and well-differentiated, naming concrete capabilities and an exact command while actively distinguishing the skill from the related OpenAI-as-MCP-tool case. Its main weakness is the absence of an explicit "Use when..." trigger clause and narrow trigger-term coverage, which cap completeness and trigger quality at 2.

Suggestions

Add an explicit 'Use when...' trigger clause (e.g., 'Use when a group should run on Codex instead of Claude') to lift completeness to 3.

Broaden trigger terms with natural user phrasings like 'switch agent provider', 'run ChatGPT/Codex as the agent', or 'use OpenAI as the planner' to improve trigger-term coverage.

Lead with the what/when pairing before the auth and command details so the trigger context is front-loaded.

Dimension	Reasoning	Score
Specificity	Lists multiple concrete capabilities ("planning, tool orchestration, MCP tools, server-side history, session resume") plus an exact activation command ("ncl groups config update --provider codex"), matching the multiple-specific-actions anchor.	3 / 3
Completeness	It clearly states what the skill does but the "when to use it" guidance is only implied; per the rubric guideline a missing explicit "Use when..." clause caps completeness at 2.	2 / 3
Trigger Term Quality	Real natural keywords exist ("Codex", "ChatGPT subscription", "OpenAI API key", "agent provider") but coverage is narrow product naming rather than broad natural variations, and no "Use when..." trigger clause is present.	2 / 3
Distinctiveness Conflict Risk	It explicitly disambiguates from the adjacent case ("Distinct from using OpenAI as an MCP tool where Claude remains the planner") and from Claude itself, giving a clear niche unlikely to trigger the wrong skill.	3 / 3
	Total	10 / 12 Passed

Validation

93%

Warnings & errors only

Checks the skill against the spec for correct structure and formatting. All validation checks must pass before discovery and implementation can be scored.

Validation — 15 / 16 Passed

Validation for skill structure

Criteria	Description	Result
relative_links	Relative link issues: 1 suspicious	Warning

	Total	15 / 16 Passed

Repository: nanocoai/nanoclaw
Path: .claude/skills/add-codex/SKILL.md
Commit: 641963c

Reviewed: 6 days ago

Table of Contents

Discovery Implementation Validation

Is this your skill?

If you maintain this skill, you can claim it as your own. Once claimed, you can manage eval scenarios, bundle related skills, attach documentation or rules, and ensure cross-agent compatibility.