guidance

Control LLM output with regex and grammars, guarantee valid JSON/XML/code generation, enforce structured formats, and build multi-step workflows with Guidance - Microsoft Research's constrained generation framework

Quality

73%

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/llm-tools/guidance/SKILL.md

Quality

Content

65%

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

Highly actionable with extensive executable code and real bundle references, but the SKILL.md is long and redundant, doing detail work that should live in the reference files. Tightening duplication and offloading patterns to references would raise conciseness and progressive disclosure.

Suggestions

Move the five "Common Patterns" and the ReAct/stateful-function examples into references/examples.md, keeping only one representative quick-start example inline.

De-duplicate repeated blocks (email regex, Anthropic backend init, generate_person) by referencing a single canonical snippet, and trim the marketing "Performance Characteristics" and "Comparison to Alternatives" sections or move them to a reference.

Add an explicit sequenced workflow with validation (e.g., constrain → generate → parse/validate output → retry on parse failure) to give workflow_clarity explicit checkpoints.

Dimension	Reasoning	Score
Conciseness	The ~560-line body is mostly efficient executable code but repeats patterns (ReAct agent and generate_person each appear twice, email regex repeated many times) and includes marketing-style prose and a comparison table that could be trimmed, matching the score-2 anchor rather than a lean score-3.	2 / 3
Actionability	It provides abundant copy-paste-ready, executable Python for regex/select/grammar/guidance-functions, backends, and common patterns, matching the score-3 anchor for fully executable guidance.	3 / 3
Workflow Clarity	There is a clear install→load→constrain→generate progression, but the patterns are presented in parallel rather than as a sequenced workflow and there are no explicit validation checkpoints or feedback loops, matching the score-2 anchor.	2 / 3
Progressive Disclosure	Three real one-level-deep references (constraints.md, backends.md, examples.md) are signaled in a See Also section, but extensive inline examples and patterns that belong in references/examples.md make the SKILL.md nearly monolithic, matching the score-2 anchor.	2 / 3
	Total	9 / 12 Passed

Description

82%

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 strong, specific description that names concrete capabilities and a distinct niche, but it lacks an explicit "Use when..." trigger clause, which caps completeness. Adding a short trigger condition would lift it to full marks.

Suggestions

Append an explicit "Use when..." clause, e.g. "Use when the user needs guaranteed valid structured output (JSON/XML/code), regex/grammar-constrained generation, or multi-step generation workflows."

Add common user phrasings like "structured output", "force valid JSON", or "constrain model output" to broaden natural trigger coverage.

Dimension	Reasoning	Score
Specificity	Lists multiple concrete actions such as "Control LLM output with regex and grammars", "guarantee valid JSON/XML/code generation", and "build multi-step workflows", matching the multi-action score-3 anchor rather than the partial score-2 anchor.	3 / 3
Completeness	It clearly answers "what" with concrete capabilities but has no "Use when..." clause or equivalent explicit trigger guidance, so completeness is capped at 2 per the rubric guideline.	2 / 3
Trigger Term Quality	Includes natural terms developers would say ("valid JSON/XML/code", "regex", "structured formats", "multi-step workflows") with good coverage, satisfying the score-3 anchor despite the somewhat jargon-ish "constrained generation".	3 / 3
Distinctiveness Conflict Risk	It targets a clear niche (constrained generation via the named Guidance framework) with distinct triggers unlikely to conflict with other skills, matching the score-3 anchor.	3 / 3
	Total	11 / 12 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 (574 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/llm-tools/guidance/SKILL.md
Commit: 52845c3

Reviewed: about 13 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.