flux-best-practices

Comprehensive guide for BFL FLUX image generation models. Covers prompting, T2I, I2I, structured JSON, hex colors, typography, multi-reference editing, and model-specific best practices for FLUX.2 and FLUX.1 families.

Quality

52%

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 ./.agents/skills/flux-best-practices/SKILL.md

Quality

Content

55%

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

The SKILL.md body is concise and well-organized with strong quick-reference material, but it leans on a rules/ bundle and AGENTS.md that do not exist in the directory, which breaks progressive disclosure and leaves the in-file guidance incomplete.

Suggestions

Ship the referenced rules/*.md files and AGENTS.md (or remove the dead links) so the progressive-disclosure structure actually resolves.

Add an explicit verification checkpoint, e.g. 'Before submitting: confirm no negative prompts, lighting is specified, and quoted text is used for any rendered words.'

Inline at least one minimal executable API snippet (model name + prompt + a parameter note) so actionability does not depend solely on the missing rule files.

Dimension	Reasoning	Score
Conciseness	The body is lean and well-paced: a one-line purpose, a tight model-selection table, six critical rules, and a single example prompt — every section earns its place and it does not explain concepts Claude already knows.	3 / 3
Actionability	It gives concrete, copy-ready building blocks (prompt-structure formula, a rules table, a model-selection table, an example prompt), but the actual detailed guidance is deferred to rule files that do not exist in the bundle, so the executable guidance in SKILL.md itself is incomplete rather than fully copy-paste ready.	2 / 3
Workflow Clarity	The 'When to Use' list and model-selection table sequence the decision of which model and prompt shape to use, but there is no explicit validation/verification step (e.g. how to check a generated prompt against the critical rules before submission), so checkpoints are only implicit.	2 / 3
Progressive Disclosure	The body references AGENTS.md and ten rules/*.md files as one-level-deep navigation, but none of these bundle files are present in the directory, so the disclosure structure is broken — the claimed detailed materials cannot actually be loaded.	1 / 3
	Total	8 / 12 Passed

Description

50%

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 a strong feature inventory for FLUX prompting but lacks an explicit 'Use when...' trigger clause and leans on technical framing over natural user language, capping completeness and trigger quality at 2.

Suggestions

Add an explicit trigger clause, e.g. 'Use when the user asks to generate or edit images with FLUX.2 or FLUX.1 models, render text in images, or match brand colors.'

Reframe capability phrases as concrete user actions ('Generate images', 'Edit existing images', 'Render quoted text') rather than the passive 'Covers...' inventory.

Disambiguate from the bfl-api skill in the description (e.g. 'For API integration see bfl-api') so triggers route correctly.

Dimension	Reasoning	Score
Specificity	Names the domain (FLUX image generation) and several specific capabilities (T2I, I2I, structured JSON, hex colors, typography, multi-reference editing), but it reads as a feature inventory ('Covers...') rather than concrete actions a user would invoke, which is closer to 'names domain and some actions' than a list of multiple specific concrete actions.	2 / 3
Completeness	It clearly answers 'what does this do' but has no explicit 'Use when...' clause or trigger guidance — the 'when' is only implied from the capability list, which the rubric caps at 2.	2 / 3
Trigger Term Quality	It surfaces relevant user-facing terms (FLUX, FLUX.2, FLUX.1, T2I, I2I, typography, hex colors) but omits the most natural phrasings a user would say ('generate an image', 'edit this image', 'render text in an image', 'make a picture'), and 'structured JSON' is technical framing rather than a user utterance.	2 / 3
Distinctiveness Conflict Risk	It is tied to a clear niche (BFL FLUX models) that is unlikely to conflict with unrelated skills, but it overlaps with the separate 'bfl-api' skill referenced in the body, so trigger-based distinction is only partially established.	2 / 3
	Total	8 / 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: 12 missing	Warning

	Total	15 / 16 Passed

Repository: calesthio/OpenMontage
Path: .agents/skills/flux-best-practices/SKILL.md
Commit: 0af32ce

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.