CtrlK
BlogDocsLog inGet started
Tessl Logo
SkillPublic

add-gc-statepoints

Add garbage-collection support to an LLVM 22 IR frontend. Covers the shadow-stack (gcroot) model for simple GCs and the statepoint/gc.relocate model for moving/relocating collectors, including the RewriteStatepointsForGC pass and StackMap section.

Quality

75%

Does it follow best practices?

Run evals on this skill

Adds up to 20 points to the overall score

View guide
SecuritybySnyk

Passed

No findings from the security scan

Fix and improve this skill with Tessl

tessl review fix ./tiles/tessl-llvm/skills/add-gc-statepoints/SKILL.md
SKILL.md
Quality
Evals
Security

Quality

Content

71%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 body is concise, actionable, and well-structured for a self-contained skill, with concrete code for both GC models. Its main weakness is workflow clarity: verification commands are listed but not framed as explicit validate-then-proceed checkpoints with error-recovery loops for fragile codegen operations.

Suggestions

Turn the verification section into an explicit sequence with a checkpoint: compile → inspect stack map → if missing, confirm setGC ran before RewriteStatepointsForGC → only proceed when the stack map is present.

Add a short error-recovery note tying common verification failures (no stack map, stale pointer use) back to the matching 'Common mistakes' item.

Tighten a few evaluative asides ('Easy to implement', 'advanced / understanding only') to keep the tone purely instructional.

DimensionReasoningScore

Conciseness

Mostly lean and efficient — step tables and code blocks earn their place with minimal padding — though a few framing sentences ('Easy to implement', 'advanced / understanding only') could be trimmed without losing clarity.

4 / 5

Actionability

Provides concrete, executable C++ and LLVM-IR snippets plus shell commands for both models, with only minor gaps (the manual-emission C++ API is left as 'prefer the pass' rather than a complete example).

4 / 5

Workflow Clarity

Steps are sequenced (Step 0 → A/B → verify → common mistakes), but the verification section lists commands without an explicit validate-then-proceed checkpoint or error-recovery loop, and there are no validation gates for the destructive/irreversible codegen operations the rubric flags.

3 / 5

Progressive Disclosure

For a self-contained single-purpose skill under 50 substantive lines with no external bundle files, the content is well-organized into clear sections (model choice, gcroot, statepoints, verification, mistakes), which satisfies the simple-skill exception for a top score.

5 / 5

Total

16

/

20

Passed

Description

66%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 highly distinct, but it omits an explicit 'Use when...' trigger clause, which caps its completeness at 3. Adding a one-sentence usage trigger would raise the overall quality.

Suggestions

Add an explicit 'Use when...' clause, e.g. 'Use when adding a garbage collector to an LLVM-based frontend and you need to mark GC-safe points and live pointers.'

Include a few natural user-facing trigger phrases like 'garbage collector', 'moving GC', or 'GC integration' alongside the technical terms.

Consider naming the target audience/state ('language runtime', 'frontend') earlier so the trigger reads more naturally.

DimensionReasoningScore

Specificity

Lists multiple concrete capabilities — adding gcroot/shadow-stack support, the statepoint/gc.relocate model, the RewriteStatepointsForGC pass, and the StackMap section — giving comprehensive coverage of the skill's scope.

5 / 5

Completeness

The 'what' is clearly and concretely stated, but there is no explicit 'Use when...' clause telling Claude when to invoke the skill, which the guidelines say caps completeness at 3.

3 / 5

Trigger Term Quality

Includes strong natural terms ('garbage-collection', 'GC-safe points', 'live pointers') but leans on technical jargon (gcroot, statepoint) rather than the everyday phrases a user would say, so a few natural synonyms are missing.

4 / 5

Distinctiveness Conflict Risk

It occupies a very specific niche — LLVM 22 IR frontend GC support — with distinct technical triggers that would not overlap with other skills.

5 / 5

Total

17

/

20

Passed

Validation

100%

Checks the skill against the spec for correct structure and formatting. All validation checks must pass before discovery and implementation can be scored.

Validation — 16 / 16 Passed

Validation for skill structure

No warnings or errors.

Repository

hbarve1/tessl-llvm

Path

tiles/tessl-llvm/skills/add-gc-statepoints/SKILL.md

Reviewed

Table of Contents

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.