CtrlK
BlogDocsLog inGet started
Tessl Logo
SkillPublic

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 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).

Quality

81%

Does it follow best practices?

Run evals on this skill

Adds up to 20 points to the overall score

View guide
SecuritybySnyk

Low

Low-risk findings worth noting

SKILL.md
Quality
Evals
Security

Quality

Content

88%Weight 40%Scale 1-5

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

An exceptionally actionable, well-sequenced install runbook: machine-applicable idempotent directives, explicit validation and error-recovery checkpoints, and lean prose that assumes reader competence. The only meaningful gaps are minor tightening opportunities and slightly better signposting/splitting of the payload manifest and external doc references.

DimensionReasoningScore

Conciseness

Dense and efficient — it assumes competence (never explains what JSON-RPC, pnpm, or a barrel is) and nearly every paragraph conveys a non-obvious constraint (vault-only credential swapping, version-pin coupling with the gateway skill). Not 5: the payload-tree description paragraph and the exchange-archive removal note could be tightened slightly.

4 / 5

Actionability

Fully executable throughout: exact nc:copy source→destination mappings, idempotent nc:append/nc:json-merge directives with the pinned version entry, concrete build and validation commands, ncl group commands, and copy-paste troubleshooting greps keyed to specific error strings. Covers the common cases end to end.

5 / 5

Workflow Clarity

Clear sequence (Pre-flight → Copy payload → Wire barrels → CLI manifest → Build → Validate → Authenticate → Use it) with a hard stop-condition in pre-flight, an already-installed check that skips to auth, an explicit validation step (provider-contract-verifier plus red-barrel registration tests), a restart-before-use checkpoint, and a symptom→cause→fix troubleshooting loop. Not 4 — the checkpoints and error-recovery feedback are explicit, not implicit.

5 / 5

Progressive Disclosure

Good structure: clear section headers, bulk payload content held in separate files surfaced via the directive listing, and a clearly linked migration doc. Not 5: the ~20-line payload manifest is inlined rather than split out, and the docs/provider-migration.md pointer plus payload contents are not explicitly signposted as navigable references.

4 / 5

Total

18

/

20

Passed

Description

70%Weight 40%Scale 1-5

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

A concrete, highly distinctive description that names specific capabilities and the exact switching command, with explicit disambiguation from related OpenAI-as-tool usage. Its main weakness is the absence of any 'Use when...' or equivalent trigger guidance, which both caps completeness and leaves natural trigger phrasings under-covered.

Suggestions

Add an explicit trigger clause, e.g. 'Use when the operator wants Codex/ChatGPT as the agent backend for a group, wants to switch a group away from Claude, or asks about codex as a provider.'

Include a few more natural user phrasings and synonyms (e.g. 'switch agent backend', 'codex CLI', 'use my ChatGPT subscription') to broaden trigger-term coverage.

DimensionReasoningScore

Specificity

Lists several concrete capabilities — 'planning, tool orchestration, MCP tools, server-side history, session resume' — plus the exact command 'ncl groups config update --provider codex' and concrete auth requirements (ChatGPT subscription or OpenAI API key). Not 5 because the enumeration names capabilities rather than comprehensively covering the actions the skill performs (install, wiring, auth walk-through).

4 / 5

Completeness

The 'what' is clear (Codex as a full agent provider with enumerated capabilities), but there is no 'Use when...' clause or equivalent explicit trigger guidance — 'alongside or instead of Claude' and 'Per-group via ncl...' describe how, not when. Per the rubric guideline, a missing explicit trigger caps this at 3.

3 / 5

Trigger Term Quality

Good natural keyword coverage: 'Codex', 'OpenAI', 'ChatGPT subscription', 'API key', 'agent provider', 'session resume'. Not 5 because common variations and phrasings a user might naturally say (e.g. 'switch agent backend', 'use GPT', 'codex CLI') are missing.

4 / 5

Distinctiveness Conflict Risk

Clear niche (Codex as a full planner/provider) with an explicit boundary against the nearest competing concept: 'Distinct from using OpenAI as an MCP tool (where Claude remains the planner)'. Minimal conflict risk.

5 / 5

Total

16

/

20

Passed

Validation

81%

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

Validation — 13 / 16 Passed

Validation for skill structure

CriteriaDescriptionResult

metadata_version

'metadata.version' is missing

Warning

relative_links

Relative link issues: 1 suspicious

Warning

referenced_paths_exist

Referenced path issues: 1 missing

Warning

Total

13

/

16

Passed

Repository

nanocoai/nanoclaw

Path

.claude/skills/add-codex/SKILL.md

Reviewed

Table of Contents

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.