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 nda-clause-taxonomygit 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/nda-clause-taxonomy)<a href="https://agentmods.dev/skills/benchflow-ai/skillsbench/nda-clause-taxonomy"><img src="https://agentmods.dev/badge/skills/benchflow-ai/skillsbench/nda-clause-taxonomy/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/nda-clause-taxonomy"><img src="https://agentmods.dev/badge/skills/benchflow-ai/skillsbench/nda-clause-taxonomy.svg" alt="Reviewed on agentmods" width="80" height="20"></a>- NVIDIA SkillSpector warn
SkillSpector: 1 finding, up to medium
These are SkillSpector’s own severities. On a checked sample its high-severity flags on skills were ~96% false positives — a documented command, a public API, a “never do X” rule — so we show them as a caution to read, not a verdict. Why →
- medium Excessive Agency · line 97 Skill enables autonomous high-impact decisions without human-in-the-loop verification. Critical operations (destructive commands, financial transactions, data deletion) should require explicit user confirmation.Fix: Add human-in-the-loop confirmation for destructive, irreversible, or high-impact operations. Never auto-execute commands that modify files, send data, or alter system state.
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 | $0.00073 | $0.02508 |
| Opus 5 | $0.00036 | $0.01254 |
| Sonnet 5 | $0.00015 | $0.00502 |
| Haiku 4.5 | $0.00007 | $0.00251 |
Grade A, and why
nda-clause-taxonomy 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 11d 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 — 143 lines — stays where its author put it; the contents beside it link to each section on GitHub.
NDA clause taxonomy
A working reference for someone reviewing a commercial NDA. Every NDA reuses the same handful of clauses; the wording varies. Use this to recognise which provision is which when the section headings don't match the clause function.
M&A NDA vs. ordinary-evaluation NDA
Before applying any review rule, identify which kind of NDA you're looking at, because conventions differ:
- M&A / transactional NDA (executed in connection with diligence on a possible acquisition or business combination). Standstills are standard — Eastland's 2018 EDGAR survey of 143 transactional NDAs found standstills in 80%, modal duration 12 months. Definition of CI typically pulls in "the existence of negotiations" (which is itself something the acquirer side often pushes back on, because it restricts the acquirer's ordinary-course communications). Survival caps tend to be longer (3-5 years) than ordinary NDAs. Residuals carve-outs are typically refused by the disclosing target, especially in pharma and biotech (Farrer & Co.).
- Ordinary business-evaluation NDA (vendor evaluation, partnership discussion, technical exchange). Standstills are unusual and a red flag. Term and survival are shorter (1-2 years term, 2-3 years survival). Residuals carve-outs more frequent on the recipient side (especially tech).
Apply rules in their proper context — a "must_be_absent" rule for standstills is appropriate for ordinary NDAs but not for M&A NDAs.
Reading order
NDAs almost always order clauses this way:
- Recitals / preamble — parties, effective date, "Authorized Purpose"
- Definition of Confidential Information
- Exceptions / carve-outs from the definition
- Use and disclosure obligations (sometimes split into "non-use" and "non-disclosure")
- Permitted recipients / Representatives
- Compelled-disclosure exception (court order, subpoena)
- Return or destruction on termination
- No-license statement
- No-warranty / "as is"
- Term and survival
- Equitable / injunctive relief
- Miscellaneous: governing law, assignment, severability, notice, counterparts, entire-agreement, waiver
- (M&A context only) Standstill, non-solicit / no-hire
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.
- 11d ago First seen · 143 lines · 73 tokens per session scan A 549b887a54dd
nda-clause-taxonomy is a skill published in the GitHub repository benchflow-ai/skillsbench (1,760 stars, last pushed 1mo ago), licensed Apache-2.0. It adds 73 tokens to every session and 2,508 once invoked, about $0.0004 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
lq-governance-playbook-benchmark
Use when benchmarking a board-level governance document against the LQ Governance Playbook — a Delegation of Authority policy, committee charter, related party transaction framework, or board terms of reference. Produces a classification table (Match / Partial Match / Below Fallback / Red Flag / Omitted) with specific…
california-small-claims
How to fill out the California Small Claims Court SC-100 form, including field mapping and required vs. optional fields.
california-sc100-form
Understand the structure and fields of the California Small Claims Court Form SC-100.
ca-small-claims-sc100
California SC-100 Small Claims Court form field mapping and filling guide.
gesellschaftsrechtliche-satzungen-agb
Für Gesellschaftsrechtliche Satzungen AGB Abgrenzung: ordnet Norm, Beweislast und Gegenargument; Ergebnis: Prüfprodukt mit Risiko und nächstem Schritt. Fachgebiet: AGB-Recht-Prüfer. Route: gesellschaftsrechtliche-satzungen-agb.
content-creator-setup
One-time onboarding for the content creator workflow — content pipeline stages, trend expiration, cross-platform cascades, heavy idea parking. After successful setup this skill is excluded from selection until the marker file is deleted.