NVIDIA/SkillEvaluator is a framework for checking and evaluating AI-agent skills, which are instruction folders that extend an agent's behavior. Skill authors and agent developers use it to validate quality and security, detect overlap, generate evaluation data, and measure skills through live agent tests. The catalogue entries are skills for running these evaluation workflows.
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 NVIDIA/SkillEvaluator --skill task-listgit clone --depth 1 https://github.com/NVIDIA/SkillEvaluatorWrote 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/nvidia/skillevaluator/task-list)<a href="https://agentmods.dev/skills/nvidia/skillevaluator/task-list"><img src="https://agentmods.dev/badge/skills/nvidia/skillevaluator/task-list/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/nvidia/skillevaluator/task-list"><img src="https://agentmods.dev/badge/skills/nvidia/skillevaluator/task-list.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 | $0.00022 | $0.00473 |
| Opus 5.5 | $0.00009 | $0.00189 |
| Sonnet 5.5 | $0.00004 | $0.00095 |
| Haiku 4.5 | $0.00002 | $0.00047 |
Grade A, and why
task-list 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 1mo 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.
What it actually says
Task List Skill
Track multi-step tasks to ensure nothing is forgotten.
When to use
- The request requires 4 or more distinct steps
- You need to perform several operations in sequence
Instructions
- Add tasks before starting multi-step work.
- Mark exactly one task in progress while working.
- Update task status as each step completes.
- Keep task text concise and action-oriented.
How to use
Use the tasks tool directly. Do NOT use run_code, execute, or write_file for task management.
1. Add tasks at the start
tasks(operation="add_multiple", content="Step one; Step two; Step three")
This clears any previous tasks. Each request starts fresh.
Examples
tasks(operation="add_multiple", content="Inspect repo; Run tests; Summarize findings")
tasks(operation="start", task_id="task_1")
tasks(operation="complete", task_id="task_1")
2. Start a task before working on it
tasks(operation="start", task_id="task_1")
3. Complete a task when done
tasks(operation="complete", task_id="task_1")
4. Check what's next
tasks(operation="next")
All operations
| Operation | Args | Description |
|---|---|---|
add_multiple |
content="t1; t2; t3" |
Add tasks (clears previous) |
start |
task_id="task_1" |
Mark in_progress |
complete |
task_id="task_1" |
Mark completed |
cancel |
task_id="task_1" |
Cancel a task |
list |
(none) | Show all tasks |
next |
(none) | Get next pending task |
Rules
- One task in_progress at a time
- Start before working, complete when done
- Do not leave tasks hanging
What ships with it
3 files beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
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.
- 1mo ago First seen · 73 lines · 22 tokens per session scan A 38f1e6e96a3a
task-list is a skill published in the GitHub repository NVIDIA/SkillEvaluator (585 stars, last pushed today), licensed Apache-2.0. It adds 22 tokens to every session and 473 once invoked, about $0.0001 per session on Opus 5.5. 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
linear
Manage Linear issues, projects, cycles, and documents via API.
task
Create, edit, and manage scheduled tasks. Tasks live at /.abtars/tasks/ — each task is a directory.
tfw-daily-task
Complete a bounded everyday request after project and purpose checks, preserving selected Trace and a checked result without a second planning approval.
project
Create or revise Obsidian Project notes for goals, decisions, open questions, and lessons. Use when the user wants a project thinking notebook or wants to close and archive one. Keep task tracking in its existing system; use only when a vault artifact is requested.
skill-authoring
Create, revise, or remove reusable procedural and interactive skills; use memorystore for facts and preferences.
working-days-get-working-days
Calculate the number of working days between two dates, optionally accounting for country-specific holidays.