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

69%

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 ./skills/prompt-engineering/guidance/SKILL.md

Quality

Content

63%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 content is highly actionable with concrete executable examples and clear multi-step patterns, but it is verbose for a SKILL.md and underutilizes its reference bundle by inlining detail that belongs in the separate files. Tightening the body and weaving inline references to the bundle files would improve both conciseness and progressive disclosure.

Suggestions

Move the detailed 'Common Patterns', 'Comparison to Alternatives', and 'Performance Characteristics' sections into the existing reference files, keeping SKILL.md as a lean overview with inline 'See references/examples.md' pointers near each topic.

Remove repetitive model-initialization lines and 'Benefits:' bullet lists that restate what the code already shows, to tighten conciseness.

Replace the non-executable JSON grammar string example with real guidance syntax (a @guidance function or gen/select-based grammar) so every code block is copy-paste ready.

Dimension	Reasoning	Score
Conciseness	The body is mostly efficient actionable code, but includes noticeable padding: repeated model initialization in every example, 'Benefits:' bullet lists restating the obvious, 'How it works' prose, and marketing sections ('GitHub Stars: 18,000+', '30-50% faster') that could be trimmed.	3 / 5
Actionability	Rich, largely copy-paste-ready Python covering common cases (regex, select, functions, ReAct, JSON, classification, extraction); the one gap is the JSON grammar example (lines 192-201) which uses non-executable string-template pseudocode rather than real guidance syntax.	4 / 5
Workflow Clarity	Multi-step patterns (ReAct with a 'Done?' checkpoint, chain-of-thought, entity extraction) are clearly sequenced, and validation is largely inherent since constraints guarantee valid output; no destructive/batch context that would require explicit feedback loops.	4 / 5
Progressive Disclosure	Section structure is clean and the three reference files (constraints.md, backends.md, examples.md) are real, but the 562-line body inlines material that could live in those references (5 patterns, comparison table, performance section) and only signals them in a final 'See Also' rather than inline near relevant sections.	3 / 5
	Total	14 / 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 specific, concrete, and clearly distinguishable, but it omits any explicit 'when to use' trigger guidance, which limits its completeness. Adding a 'Use when...' clause with natural user phrasings would raise both completeness and trigger-term quality.

Suggestions

Add an explicit 'Use when...' clause (e.g., 'Use when you need guaranteed valid JSON/XML from an LLM, regex-constrained output, or structured multi-step generation') to satisfy the completeness 'when' requirement.

Include more natural user-facing synonyms (e.g., 'structured output', 'constrain LLM responses', 'force valid JSON') alongside the technical terms to improve trigger-term quality.

Drop the marketing attribution ('Microsoft Research's constrained generation framework') from the description — it adds length without aiding trigger matching.

Dimension	Reasoning	Score
Specificity	Lists multiple concrete actions — 'Control LLM output with regex and grammars', 'guarantee valid JSON/XML/code generation', 'enforce structured formats', 'build multi-step workflows' — giving comprehensive coverage of the skill's capabilities.	5 / 5
Completeness	The 'what' is clearly and concretely stated, but there is no 'Use when...' clause or equivalent explicit trigger guidance for when Claude should invoke it, which caps completeness at 3 per the rubric guideline.	3 / 5
Trigger Term Quality	Good natural keywords (JSON, XML, regex, structured formats, workflows) but leans technical with 'grammars' and 'constrained generation framework', missing a few common user phrasings; not quite comprehensive enough for a 5.	4 / 5
Distinctiveness Conflict Risk	Names a specific framework ('Guidance - Microsoft Research's constrained generation framework') with distinct triggers (regex, grammars, constrained generation), giving it a clear niche with minimal conflict risk.	5 / 5
	Total	17 / 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 (573 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: OpenLAIR/dr-claw
Path: skills/prompt-engineering/guidance/SKILL.md
Commit: 4b35756

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.