tri-ai-collaboration

Complete development workflows where Claude writes the code while Gemini and Codex provide research, planning, reviews, and different perspectives. Claude remains the main developer. Use for complex projects requiring expert planning and multi-perspective reviews.

Quality

64%

Does it follow best practices?

Run evals on this skill

Adds up to 20 points to the overall score

View guide

Securityby

Low

Low-risk findings worth noting

Fix and improve this skill with Tessl

tessl review fix ./skills/tri-ai-collaboration/SKILL.md

Quality

Content

38%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.

The skill delivers a genuinely useful decision matrix and concrete CLI syntax for orchestrating Gemini/Codex as advisors, but buries them in massive redundancy and an internal contradiction: the advisory premise ('Claude writes the code') is violated by its own later phases where Codex and Gemini implement code. It needs to be cut to a fraction of its length with worked examples moved to reference files.

Suggestions

Collapse the three overlapping end-to-end examples (OAuth, notification system, SaaS dashboard) into one reference file, keeping only the decision matrix, model-selection tiers, and a single condensed quick-start workflow in SKILL.md.

Resolve the core contradiction: either keep the 'Claude writes all code' advisory model and rewrite Phases 3–7 so Gemini/Codex only advise, or reframe the skill honestly as delegated parallel execution.

Add an explicit validation checkpoint before the batch git/deploy phase (e.g. verify tests pass and review diffs before commits/PRs are created), and fix the broken bash in Pattern 3's loop.

Dimension	Reasoning	Score
Conciseness	The 1,167-line body is heavily padded: role definitions are repeated three times ('Role Clarity', 'The Power of Three', 'Summary'), and three near-duplicate end-to-end examples (OAuth, notification system, SaaS dashboard) each re-implement the same 7-phase workflow. Marketing fluff like 'Ready to build at superhuman speed?' and '10-50x faster development' adds nothing.	2 / 5
Actionability	Concrete CLI invocations are given (e.g. 'gemini --yolo --output-format json', 'codex exec -m o3 --json --dangerously-bypass-approvals-and-sandbox'), but many code blocks are stubs ('// ...', "// Claude's implementation", 'rest of improved implementation') and Pattern 3 contains a syntax error ('echo "Quality threshold met!break' is missing its closing quote).	3 / 5
Workflow Clarity	A clear 7-phase sequence exists with a test-fail→fix→re-run loop in Phase 4, but Phase 7's batch commit/PR/deploy operations have no validation checkpoint before release, and Phases 3–7 contradict the skill's stated premise by having Codex/Gemini write code directly ('Create all files and implement completely') after the body insists 'Claude is the one writing and editing code'.	3 / 5
Progressive Disclosure	No references/, scripts/, or assets/ files exist; the SKILL.md is a monolithic 1,100+ line document inlining three complete worked examples that clearly belong in separate reference files. 'Related Skills' points to files like CLAUDE-CODEX-INTEGRATION.md that are not present in any bundle.	2 / 5
	Total	10 / 20 Passed

Description

75%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.

The description is well-constructed: it states the division of roles concretely, keeps Claude's central role explicit, and includes a clear 'Use for...' trigger clause. Minor improvements would make the trigger terms more natural and vary the keywords users might actually say.

Dimension	Reasoning	Score
Specificity	Names the domain (development workflows with AI advisors) and several concrete actions: 'Claude writes the code while Gemini and Codex provide research, planning, reviews, and different perspectives'. Not 5 because it does not comprehensively cover the orchestration capabilities the body describes.	4 / 5
Completeness	Both parts are explicit: the 'what' ('Complete development workflows where Claude writes the code while Gemini and Codex provide...') and the 'when' ('Use for complex projects requiring expert planning and multi-perspective reviews'). Not 5 because the when-clause trigger is abstract — 'complex projects' is less concrete than user-sayable phrases.	4 / 5
Trigger Term Quality	Good natural keywords: 'Gemini', 'Codex', 'research', 'planning', 'reviews', 'multi-perspective reviews', 'complex projects'. Not 5 because common user phrasings like 'AI collaboration' or 'work with multiple AIs' are absent.	4 / 5
Distinctiveness Conflict Risk	Naming Gemini and Codex gives it a clear niche that is unlikely to trigger for unrelated skills. Not 5 because it still overlaps with generic multi-AI orchestration or delegation skills.	4 / 5
	Total	16 / 20 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
skill_md_line_count	SKILL.md is long (1167 lines); consider splitting into references/ and linking	Warning

	Total	15 / 16 Passed

Repository: fernandezbaptiste/Skrillz
Path: skills/tri-ai-collaboration/SKILL.md
Commit: 93ed392

Reviewed: 14 hours 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.