CtrlK
BlogDocsLog inGet started
Tessl Logo

tooluniverse-codex-plugin

Install, set up, verify, update, pin, uninstall, or troubleshoot the ToolUniverse plugin on OpenAI Codex. ALWAYS consult this skill for any of those — don't answer from memory, because the exact marketplace name (mims-harvard/ToolUniverse), the "codex plugin marketplace add" then "codex plugin add -m tooluniverse" flow, Codex's startup auto-upgrade behavior, the uvx tooluniverse MCP server, and the API-key env vars are easy to get wrong. Use it whenever someone wants to get ToolUniverse (or "the 1000+ scientific tools" / "the harvard tools") working on Codex, says the Codex plugin or its tools/skills won't load, hits a uvx or MCP-server startup error, asks how Codex updates it, wants to pin or remove it, or finds it running an old tool version — even if they never say the word "plugin". Not for the Claude Code plugin (use tooluniverse-claude-code-plugin), for running research with the tools, or for authoring new tools or skills.

75

Quality

94%

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

The canonical home for this skill is tooluniverse-codex-plugin in mims-harvard/ToolUniverse

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.

A highly actionable, well-sequenced skill body: copy-paste commands throughout, explicit verification steps, a symptom-indexed troubleshooting table, and an idempotent diagnosis script with feedback loops. The main weaknesses are repeated fix content across the Updates, Troubleshooting, and Diagnosis sections, and the maintainer-focused material being inlined in a user-facing skill.

Suggestions

Consolidate the repeated fixes: 'uv cache clean tooluniverse', the uv-tool shadowing issue, and the owner/repo-vs-path guidance each appear in two or three sections — pick one canonical location (e.g. Troubleshooting) and cross-reference it from Updates and the diagnosis script.

Move the 'For plugin maintainers' section into a separate reference file (e.g. MAINTAINERS.md) since it addresses a different audience than the user-facing install flow, keeping SKILL.md as a lean overview.

Consider moving the API-key env var examples into the referenced API_KEYS_REFERENCE.md material or a small reference file, keeping only the export-pattern pointer in SKILL.md.

DimensionReasoningScore

Conciseness

The body is commands-first with no explanations of concepts Claude already knows, but several fixes appear in multiple places: 'uv cache clean tooluniverse' appears in Updates, the Troubleshooting table, and diagnosis step 7, and the owner/repo-vs-local-path point recurs in four sections. This is 'efficient with minor instances that could be trimmed' (4) rather than 5 ('every token earns its place') — the redundancy is noticeable, though each instance serves a different entry path (linear read vs. symptom lookup).

4 / 5

Actionability

Every section provides copy-paste-ready commands with expected outputs ('expect: tooluniverse (enabled)'), including a 7-step ordered, safe, idempotent diagnosis script with explicit FIX actions per failure. This matches the 5 anchor: fully executable guidance covering the common cases.

5 / 5

Workflow Clarity

The sequence is clear (prerequisites check → two-command install → restart → verify → use) with an explicit validation checkpoint ('codex plugin list' expecting 'tooluniverse (enabled)'), a symptom→fix troubleshooting table, and a feedback loop ('apply the FIX for whatever fails, then restart Codex'). This matches the 5 anchor's validate → fix → retry pattern.

5 / 5

Progressive Disclosure

No bundle files exist, and the body's only outward reference ('the bundled setup-tooluniverse skill → API_KEYS_REFERENCE.md') is one level deep and clearly signaled. However, the 'For plugin maintainers' section targets a different audience and would sit better in a separate reference file, and the ~160-line body carries material that could be split — 'good structure with minor organization gaps' (4) rather than an ideally split overview (5).

4 / 5

Total

18

/

20

Passed

Description

100%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.

An exemplary description: concrete capability list, precise commands and names, comprehensive natural trigger phrasing including symptom-based and synonym-based variants, and explicit exclusions that separate it from the sibling Claude Code plugin skill. Its length is dense with trigger-relevant specifics rather than padding.

DimensionReasoningScore

Specificity

The description enumerates multiple concrete actions — 'Install, set up, verify, update, pin, uninstall, or troubleshoot' — and backs them with exact specifics (the 'mims-harvard/ToolUniverse' marketplace name, the 'codex plugin marketplace add' then 'codex plugin add -m tooluniverse' flow, the uvx tooluniverse MCP server, API-key env vars). This matches the 5 anchor (multiple specific concrete actions, comprehensive coverage) rather than 4, since no capability gap is apparent.

5 / 5

Completeness

The first sentence answers 'what' concretely, and 'Use it whenever someone wants to get ToolUniverse... working on Codex, says..., hits..., asks..., wants..., or finds...' provides explicit, concrete trigger guidance. This is exactly the 5 anchor pattern (clear what AND when with concrete trigger phrases), not 4 where the 'when' could be more specific.

5 / 5

Trigger Term Quality

It covers natural user phrasings comprehensively, including synonyms ('the 1000+ scientific tools' / 'the harvard tools'), symptom statements ('says the Codex plugin or its tools/skills won't load', 'hits a uvx or MCP-server startup error', 'finds it running an old tool version'), and the explicit note that users may 'never say the word "plugin"'. No natural trigger phrasing category is missing.

5 / 5

Distinctiveness Conflict Risk

It carves out a clear niche (ToolUniverse on OpenAI Codex) and explicitly disambiguates against the nearest sibling — 'Not for the Claude Code plugin (use tooluniverse-claude-code-plugin), for running research with the tools, or for authoring new tools or skills' — minimizing wrong-skill triggering.

5 / 5

Total

20

/

20

Passed

Validation

93%

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

CriteriaDescriptionResult

referenced_paths_exist

Referenced path issues: 1 missing

Warning

Total

15

/

16

Passed

Repository
mims-harvard/ToolUniverse
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.