add-intrinsic

Add a new LLVM IR intrinsic in LLVM 22. Covers TableGen definition, header regeneration, IRBuilder call-site usage, verifier rules, and lit testing.

Quality

80%

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 ./tiles/tessl-llvm/skills/add-intrinsic/SKILL.md

Quality

Content

77%Scale 1-3

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

This is a high-quality, actionable skill that provides a comprehensive step-by-step guide for adding an LLVM IR intrinsic. Its greatest strengths are the fully executable code examples at every step and the clear sequential workflow with validation checkpoints. The main weakness is that it's quite long for a SKILL.md file, with reference tables and dual-path explanations (GlobalISel + SelectionDAG) that could be split into separate files for better progressive disclosure.

Suggestions

Move the properties table and type tokens table to a separate REFERENCE.md file, linking to it from the main skill with a brief summary

Consider splitting the target-specific lowering (Step 3) into a separate LOWERING.md since it covers two distinct paths (GlobalISel and SelectionDAG) and is optional for generic intrinsics

Dimension	Reasoning	Score
Conciseness	The skill is mostly efficient and covers a complex multi-step process, but includes some unnecessary explanation (e.g., the full type tokens table, the common properties table could be trimmed or referenced externally). The tables are useful reference material but make the document quite long. It respects Claude's intelligence in most places but could be tighter.	2 / 3
Actionability	Every step includes fully executable, copy-paste-ready code examples — TableGen definitions, CMake commands, C++ lowering code, IRBuilder usage, Verifier checks, lit tests, and InstCombine folds. The examples are concrete and specific with real LLVM 22 API calls.	3 / 3
Workflow Clarity	The 8-step workflow (Steps 0-7) is clearly sequenced with logical progression from definition to testing. Validation is explicitly addressed: Step 2 includes regeneration verification, Step 5 covers verifier rules, Step 6 includes running verify passes before optimization passes, and the 'Common mistakes' section serves as a validation checklist. The note about rebuilding intrinsics_gen acts as a critical checkpoint.	3 / 3
Progressive Disclosure	The content is well-structured with clear headers and a logical flow, but it's a monolithic document with substantial inline reference material (type tokens table, properties table, both GlobalISel and SelectionDAG paths) that could be split into separate reference files. For a skill of this length (~200 lines), some content would benefit from being externalized with clear links.	2 / 3
	Total	10 / 12 Passed

Description

82%Scale 1-3

Based on the skill's description, can an agent find and select it at the right time? Clear, specific descriptions lead to better discovery.

This is a strong, highly specific description targeting a narrow technical domain. It excels at listing concrete actions and using domain-appropriate trigger terms. Its main weakness is the absence of an explicit 'Use when...' clause, which would help Claude know exactly when to select this skill.

Suggestions

Add an explicit 'Use when...' clause, e.g., 'Use when the user wants to add, define, or implement a new LLVM IR intrinsic, or asks about TableGen intrinsic definitions in LLVM 22.'

Dimension	Reasoning	Score
Specificity	Lists multiple specific concrete actions: TableGen definition, header regeneration, IRBuilder call-site usage, verifier rules, and lit testing. These are precise, actionable steps rather than vague language.	3 / 3
Completeness	Clearly answers 'what does this do' (add a new LLVM IR intrinsic with specific steps), but lacks an explicit 'Use when...' clause or equivalent trigger guidance, which caps this at 2 per the rubric.	2 / 3
Trigger Term Quality	Includes highly specific natural keywords a developer would use: 'LLVM IR intrinsic', 'LLVM 22', 'TableGen', 'IRBuilder', 'verifier rules', 'lit testing'. These are the exact terms someone working in this domain would mention.	3 / 3
Distinctiveness Conflict Risk	Extremely niche and specific to LLVM 22 IR intrinsic development. The combination of TableGen, IRBuilder, verifier rules, and lit testing makes it virtually impossible to conflict with other skills.	3 / 3
	Total	11 / 12 Passed

Validation

100%

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 — 11 / 11 Passed

Validation for skill structure

No warnings or errors.

Repository: hbarve1/tessl-llvm
Path: tiles/tessl-llvm/skills/add-intrinsic/SKILL.md
Commit: a251bde

Reviewed: 6 days 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.