SkillsBench is a benchmark for measuring how effectively AI agents use modular skills—folders containing instructions, scripts, and resources—to complete specialized tasks. It helps researchers and developers evaluate both skill quality and agent behavior, including tasks that require combining multiple skills. The catalogue’s skills and instructions are evaluated as part of this workflow.
Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx skills add benchflow-ai/skillsbench --skill geometric-layout-repairgit clone --depth 1 https://github.com/benchflow-ai/skillsbenchWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/skills/benchflow-ai/skillsbench/geometric-layout-repair)<a href="https://agentmods.dev/skills/benchflow-ai/skillsbench/geometric-layout-repair"><img src="https://agentmods.dev/badge/skills/benchflow-ai/skillsbench/geometric-layout-repair/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/skills/benchflow-ai/skillsbench/geometric-layout-repair"><img src="https://agentmods.dev/badge/skills/benchflow-ai/skillsbench/geometric-layout-repair.svg" alt="Reviewed on agentmods" width="80" height="20"></a>- NVIDIA SkillSpector pass
What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00029 | $0.01057 |
| Opus 5 | $0.00015 | $0.00528 |
| Sonnet 5 | $0.00006 | $0.00211 |
| Haiku 4.5 | $0.00003 | $0.00106 |
Grade A, and why
geometric-layout-repair scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 12d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 74 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Use this skill when the deliverable is a repaired plan-view layout and a modified CAD/DXF file, not a rendered drawing.
Representation
Use simple geometry that deterministic tests can verify:
room.polygon: ordered room boundary points.door: id, clear width, swing type, and opening segment.fixtures: id, type, bounding box, and any fixture-specific accessibility metadata.grab_bars: id, type, length, and segment endpoints.turning_space: circle type, diameter, and center.
When a repaired DXF is required, keep the JSON and DXF synchronized. The JSON is the structured explanation; the DXF is the CAD deliverable.
Minimal-Change Strategy
- Start from the extracted original layout.
- List the exact rule failures.
- Try local edits in this order:
- Adjust metadata that is already implied by the plan, such as door swing direction.
- Move a fixture a small distance to meet centerline or clearance rules.
- Reposition a turning circle within usable floor area.
- Adjust a nearby fixture locally when the turning circle cannot fit otherwise.
- Expand the room boundary only if the task explicitly allows it or no local compliant repair exists.
- Keep unchanged fixture IDs and protected fixtures stable.
- After each edit, recompute clearances and containment.
DXF Repair Output
- Start from the input DXF so the original architectural context is preserved.
- Add or update semantic repair layers for the final geometry, such as
REPAIR-ROOM,REPAIR-DOOR,REPAIR-WC,REPAIR-LAV,REPAIR-TUB,REPAIR-GRABBAR, andREPAIR-CLEARANCE. - Overlay repair layers are an acceptable final CAD repair representation when the verifier needs machine-checkable geometry. Do not spend time deleting or rewriting the original CAD layers unless the task explicitly requires destructive source-layer editing.
- Write room and fixture boundaries as closed lightweight polylines.
- Write grab bars as line segments with endpoints matching the repaired layout.
- Write the turning circle as a
CIRCLEentity with radiusdiameter / 2. - Save the repaired CAD file to the requested output path, usually
/root/output/repaired_plan.dxf. - Finish the required JSON and DXF outputs before producing optional visual previews or exploratory artifacts.
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 12d ago First seen · 74 lines · 29 tokens per session scan A 9fc46125c193
geometric-layout-repair is a skill published in the GitHub repository benchflow-ai/skillsbench (1,764 stars, last pushed 1mo ago), licensed Apache-2.0. It adds 29 tokens to every session and 1,057 once invoked, about $0.0001 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.
Other skills, from other repositories
Visual Regression Review
Compare browser-rendered UI state against expected layout, content, and responsive behavior.
technical-illustration-design
Design principles and techniques for creating technical exploded-view diagrams and engineering documentation posters.
anthropic-brand-colors
Anthropic's official brand color palette, design tokens, and typography standards for consistent brand application.
anthropic-brand-tokens
Anthropic official brand color palette, typography standards, and design tokens for consistent brand application.
pillow-technical-diagrams
Create technical poster and diagram images using Pillow (PIL), including shapes, text, lines, and layered composition.
pillow-technical-drawing
Creating technical diagrams and exploded-view illustrations using Python Pillow with geometric primitives and text annotations.