himank-test/tessl-llvm

LLVM 22.x tile for building compilers, language runtimes, and out-of-tree tooling

1.23x

Quality

83%

Does it follow best practices?

Impact

96%

1.23x

Average score across 5 eval scenarios

Securityby

Passed

No known issues

Skills Audit Report — himank-test/tessl-llvm

Name: himank-test/tessl-llvm
Rating: 88.21 (1 reviews)
Author: himank-test

Date: 2026-04-08
Tile version at time of audit: 0.4.0

Methodology

Compared every file in docs/ against the skills registered in tile.json. Any doc without a corresponding skill was flagged as a gap. Additionally, common LLVM tasks not yet covered by any doc or skill were noted.

Gaps found at audit time

1. Doc-backed gaps (doc existed, skill did not)

Doc file	Missing skill ID	Added?
`docs/exception-handling.md`	`add-exception-handling`	✅
`docs/gc-statepoints.md`	`add-gc-statepoints`	✅
`docs/attributes-metadata.md`	`add-attributes-metadata`	✅
`docs/lto.md`	`add-lto`	✅
`docs/alias-analysis.md`	`add-alias-analysis`	✅

2. Additional gaps (no doc or skill existed)

Topic	Missing skill ID	Added?
Sanitizer instrumentation (ASan, UBSan, TSan)	`add-sanitizer`	✅
Auto-vectorization and loop hints	`add-vectorization-hint`	✅
Struct / union / tuple type lowering	`lower-struct-types`	✅

Skill inventory after audit

Skill ID	Skill file	Backed by doc?
`add-npm-pass`	`skills/tessl-llvm/SKILL.md`	`docs/new-pass-manager.md`
`out-of-tree-setup`	`skills/out-of-tree-setup/SKILL.md`	`docs/out-of-tree.md`
`add-intrinsic`	`skills/add-intrinsic/SKILL.md`	`docs/ir-types.md`
`version-sync`	`skills/version-sync/SKILL.md`	`docs/version-notes.md`
`new-target-backend`	`skills/new-target/SKILL.md`	`docs/codegen.md`
`frontend-to-ir`	`skills/frontend-to-ir/SKILL.md`	`docs/frontend-to-ir.md`
`add-debug-info`	`skills/add-debug-info/SKILL.md`	`docs/debug-info.md`
`jit-setup`	`skills/jit-setup/SKILL.md`	`docs/jit.md`
`lit-filecheck`	`skills/lit-filecheck/SKILL.md`	(standalone skill)
`add-calling-convention`	`skills/add-calling-convention/SKILL.md`	`docs/calling-conventions.md`
`add-exception-handling`	`skills/add-exception-handling/SKILL.md`	`docs/exception-handling.md`
`add-gc-statepoints`	`skills/add-gc-statepoints/SKILL.md`	`docs/gc-statepoints.md`
`add-attributes-metadata`	`skills/add-attributes-metadata/SKILL.md`	`docs/attributes-metadata.md`
`add-lto`	`skills/add-lto/SKILL.md`	`docs/lto.md`
`add-alias-analysis`	`skills/add-alias-analysis/SKILL.md`	`docs/alias-analysis.md`
`add-sanitizer`	`skills/add-sanitizer/SKILL.md`	(no doc yet)
`add-vectorization-hint`	`skills/add-vectorization-hint/SKILL.md`	(no doc yet)
`lower-struct-types`	`skills/lower-struct-types/SKILL.md`	(no doc yet)

Total skills: 18

Remaining doc-only content (no skill yet)

These docs exist but cover concepts that are typically referenced rather than acted on step-by-step, so a procedural skill was not added. Revisit if users request guided workflows for these topics.

Doc file	Notes
`docs/tablegen.md`	Covered implicitly by `new-target-backend`; could become a standalone `write-tablegen-defs` skill
`docs/ir-types.md`	General reference; covered by `frontend-to-ir` and `lower-struct-types`
`docs/codegen.md`	Deep-dive reference; covered by `new-target-backend`
`docs/index.md`	Tile landing page — not a skill target

Suggested future additions

Skill ID	Description
`write-tablegen-defs`	Step-by-step: define registers, instructions, and calling conventions in TableGen for a new target
`add-sanitizer-runtime`	Write a custom sanitizer runtime (shadow memory, callbacks) rather than using Clang's
`emit-wasm-target`	LLVM 22 WebAssembly target specifics: emscripten triple, exception model, wasm-specific intrinsics
`add-profile-guided-optimization`	Emit PGO instrumentation, collect profiles, and feed them back into the optimization pipeline

docs

evals

rules

skills

AUDIT.md

tile.json