fireworks-ai-inference

Fast inference and fine-tuning platform with serverless and on-demand GPU deployments. OpenAI-compatible API for chat completions, embeddings, function calling, vision, and structured output. Supports SFT, DPO, and RL fine-tuning. SOC2 + HIPAA compliant.

Quality

66%

Does it follow best practices?

Run evals on this skill

Adds up to 20 points to the overall score

View guide

Securityby

Critical

Do not install without reviewing

Fix and improve this skill with Tessl

tessl review fix ./backend/cli/skills/cloud-compute/fireworks-ai/SKILL.md

Quality

Content

61%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 body is highly actionable with executable examples across the platform's surface area, but it is verbose, lacks validation checkpoints for batch/destructive workflows, and inlines reference material that should be split into separate files.

Suggestions

Add explicit validation checkpoints to batch/destructive workflows: after creating a fine-tuning job poll `state` until non-pending before deploying, verify a deployment is READY before querying it, and validate batch JSONL format before upload.

Split bulk reference material (full CLI commands, model catalog, fine-tuning API details, embeddings model list) into separate referenced files (e.g. cli.md, models.md, fine-tuning.md) and link to them from SKILL.md.

De-duplicate the OpenAI client setup — show it once in Quick Start and reference it elsewhere — to reduce token cost.

Dimension	Reasoning	Score
Conciseness	Mostly efficient with executable code, but the ~670-line body repeats the OpenAI client setup across Quick Start, Inference, and OpenAI Compatibility sections and inlines large reference tables that could be trimmed.	3 / 5
Actionability	Nearly every section gives copy-paste-ready, executable code or commands covering the common cases — chat, streaming, tools, structured output, vision, fine-tuning jobs, deployments, embeddings, and CLI.	5 / 5
Workflow Clarity	Sequences are present for fine-tuning, deployments, and batch inference, but these batch/destructive operations lack validation checkpoints (e.g. verify job state before querying, confirm deployment readiness, validate batch file format), which per the rubric caps workflow clarity at 3.	3 / 5
Progressive Disclosure	No bundle files exist and all reference content — full CLI reference, model catalog, fine-tuning detail, embeddings list — is inlined in one monolithic SKILL.md rather than split into clearly signaled one-level-deep files.	2 / 5
	Total	13 / 20 Passed

Description

71%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 specific and capability-rich, cleanly conveying what Fireworks AI does. Its main weakness is the missing 'when to use' trigger guidance, which caps completeness and slightly weakens trigger-term quality.

Suggestions

Add an explicit 'Use when...' clause naming natural trigger phrases (e.g. 'Use when you need fast OpenAI-compatible inference for open models like Llama, Qwen, or DeepSeek, or when fine-tuning via SFT/DPO/RL without managing GPUs').

Include common user-facing synonyms such as 'LLM API', 'open-model hosting', and 'GPU inference' alongside the capability list.

Tighten the compliance and capability enumeration so a user's natural request wording maps directly to a trigger.

Dimension	Reasoning	Score
Specificity	Lists multiple concrete capabilities — 'chat completions, embeddings, function calling, vision, and structured output' plus 'SFT, DPO, and RL fine-tuning' and 'serverless and on-demand GPU deployments' — giving comprehensive coverage of the platform's actions.	5 / 5
Completeness	It clearly states 'what' the platform does but contains no 'Use when...' or equivalent trigger guidance for when to invoke the skill, which per the rubric caps completeness at 3.	3 / 5
Trigger Term Quality	Good keyword coverage with natural terms like inference, fine-tuning, chat completions, embeddings, and vision, but it lacks the common phrasings a user would actually say ('OpenAI-compatible', 'LLM API') and has no explicit trigger phrasing.	4 / 5
Distinctiveness Conflict Risk	The Fireworks AI niche and specific capability list make it mostly distinct, though the absence of explicit triggers means it could still overlap with other cloud-compute / inference skills.	4 / 5
	Total	16 / 20 Passed

Validation

87%

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 — 14 / 16 Passed

Validation for skill structure

Criteria	Description	Result
skill_md_line_count	SKILL.md is long (679 lines); consider splitting into references/ and linking	Warning
frontmatter_unknown_keys	Unknown frontmatter key(s) found; consider removing or moving to metadata	Warning

	Total	14 / 16 Passed

Repository: synthetic-sciences/openscience
Path: backend/cli/skills/cloud-compute/fireworks-ai/SKILL.md
Commit: 3a6c3a9

Reviewed: 1 day 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.