Skip to content

feat: improve skill score for tool-description-error-tracking#26

Open
yogesh-tessl wants to merge 1 commit into
llnl:masterfrom
yogesh-tessl:improve/skill-review-optimization
Open

feat: improve skill score for tool-description-error-tracking#26
yogesh-tessl wants to merge 1 commit into
llnl:masterfrom
yogesh-tessl:improve/skill-review-optimization

Conversation

@yogesh-tessl

Copy link
Copy Markdown

Hey @ilagunap 👋

a floating-point checker built specifically for HPC is something I hadn't seen before. The fact that you only need to swap out the compiler invocation to get it working is a really smart design choice. Clean docs too

ran your skills through tessl skill review at work and found some targeted improvements for the tool-description-error-tracking skill. Here's the full before/after:

Skill Before After Change
tool-description-error-tracking 26% 86% +60%
Changes made to tool-description-error-tracking
  • Expanded frontmatter description - replaced the generic "Guide for understanding the tool's functionality" with a specific description naming FPChecker, its LLVM plugin, and concrete actions (compile, run, interpret reports), plus a "Use when..." clause with natural trigger terms
  • Restructured content into a clear workflow - replaced loosely organized sections with a numbered step-by-step process: ensure PATH → compile → run → read reports
  • Removed unnecessary verbosity - trimmed explanations of concepts Claude already knows while preserving the valuable domain-specific details (FP32 vs FP64 comparison, ODR linkage, _FPC_FP32_CALCULATE_ERROR_ internals)
  • Fixed typos - corrected "firdt" → "first", "variablea" → "variables", "oprand" → "operand", "instrution" → "instruction", "oprations" → "operations", "incude" → "include"
  • Replaced hardcoded paths - swapped user-specific paths with generic $PROJECT_ROOT and $INSTALL_PATH placeholders
  • Consolidated key source files - organized the four critical files with brief role descriptions for quick reference

quick honest disclosure. I work at https://github.com/tesslio where we build tooling around skills like these. Not a pitch, just saw room for improvement and wanted to contribute.

If you want to self-improve your skills, or define your own scenarios to pressure test, just ask your agent (Claude Code, Codex, etc.) to evaluate and optimize your skill with Tessl. Ping me @yogesh-tessl, if you hit any snags.

Hey @ilagunap 👋

I ran your skills through `tessl skill review` at work and found some targeted improvements for the `tool-description-error-tracking` skill. Here's the full before/after:

| Skill | Before | After | Change |
|-------|--------|-------|--------|
| tool-description-error-tracking | 26% | 86% | +60% |
| build-the-tool | 56% | 56% | — |
| html-reports-description | 35% | 35% | — |
| write-tests | 31% | 31% | — |

<details>
<summary>Changes made to <code>tool-description-error-tracking</code></summary>

- **Expanded frontmatter description** — replaced the generic "Guide for understanding the tool's functionality" with a specific description naming FPChecker, its LLVM plugin, and concrete actions (compile, run, interpret reports), plus a "Use when..." clause with natural trigger terms
- **Restructured content into a clear workflow** — replaced loosely organized sections with a numbered step-by-step process: ensure PATH → compile → run → read reports
- **Removed unnecessary verbosity** — trimmed explanations of concepts Claude already knows while preserving the valuable domain-specific details (FP32 vs FP64 comparison, ODR linkage, `_FPC_FP32_CALCULATE_ERROR_` internals)
- **Fixed typos** — corrected "firdt" → "first", "variablea" → "variables", "oprand" → "operand", "instrution" → "instruction", "oprations" → "operations", "incude" → "include"
- **Replaced hardcoded paths** — swapped user-specific paths with generic `$PROJECT_ROOT` and `$INSTALL_PATH` placeholders
- **Consolidated key source files** — organized the four critical files with brief role descriptions for quick reference

</details>

I also stress-tested your `build-the-tool` skill against a few real-world task evals and it held up really well on the cmake + clang-19 build workflow with conda environment setup. Kudos for that.

Honest disclosure — I work at @tesslio where we build tooling around skills like these. Not a pitch — just saw room for improvement and wanted to contribute.

Want to self-improve your skills? Just point your agent (Claude Code, Codex, etc.) at [this Tessl guide](https://docs.tessl.io/evaluate/optimize-a-skill-using-best-practices) and ask it to optimize your skill. Ping me — [@yogesh-tessl](https://github.com/yogesh-tessl) — if you hit any snags.

Thanks in advance 🙏
@yogesh-tessl

Copy link
Copy Markdown
Author

hey @ilagunap, just a gentle nudge in case this got buried, let me know if you need any changes!
No rush at all if this is in a review pipeline.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant