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 paired-reversal-cancellationgit 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/paired-reversal-cancellation)<a href="https://agentmods.dev/skills/benchflow-ai/skillsbench/paired-reversal-cancellation"><img src="https://agentmods.dev/badge/skills/benchflow-ai/skillsbench/paired-reversal-cancellation/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/paired-reversal-cancellation"><img src="https://agentmods.dev/badge/skills/benchflow-ai/skillsbench/paired-reversal-cancellation.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 | $0.00112 | $0.01391 |
| Opus 5 | $0.00056 | $0.00696 |
| Sonnet 5 | $0.00022 | $0.00278 |
| Haiku 4.5 | $0.00011 | $0.00139 |
Grade A, and why
paired-reversal-cancellation 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 13d 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 — 124 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Paired reversal cancellation
GL batch tapes that support corrections distinguish two reversal flavors:
| Flavor | Shape on tape | Effect on balance |
|---|---|---|
| Unpaired RV | RV row, ref-trace blank |
Posts via the standard RV rule for the account type (e.g. sign-flip on a deposit). |
| Paired RV | RV row, ref-trace = an earlier trace, within the reversal window |
The RV row AND the row it references are BOTH suppressed; both reported as CANCEL rejects; net delta to the balance is zero. |
The intent of a paired RV is "this earlier posting was a mistake — erase
it entirely." If only the RV is suppressed, the original still posts and
the running balance ends up off by 2 × magnitude (the unwanted original
plus the absence of the reversal that would have offset it).
The reversal settlement window
A reference-trace match is necessary but not sufficient to cancel a pair. Shops bound how far back a reversal may reach: a paired RV only cancels both legs when the RV's value-date and the referenced posting's value-date are within the reversal settlement window of each other. The window for this ledger is 2 days (inclusive).
If the two value-dates are more than 2 days apart, the reference is stale: the RV is not treated as a cancellation. It falls through to the standard RV posting rule for the account type (e.g. a sign-flip / magnitude subtract on a deposit), the referenced original also posts normally, and neither leg is rejected. So the discriminator has two parts — a matching ref-trace and a date gap inside the window.
Compute the gap from the two value-dates as a true day count (e.g. via
FUNCTION INTEGER-OF-DATE on each YYYYMMDD, then the absolute difference) —
not by subtracting the raw YYYYMMDD integers, which is wrong across month and
year boundaries.
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.
- 13d ago First seen · 124 lines · 112 tokens per session scan A dc2a219c6cc7
paired-reversal-cancellation is a skill published in the GitHub repository benchflow-ai/skillsbench (1,764 stars, last pushed 1mo ago), licensed Apache-2.0. It adds 112 tokens to every session and 1,391 once invoked, about $0.0006 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
portfolio
Cross-chain DeFi portfolio discovery, rebalancing suggestions, and NEAR Intent construction. Activates when the user pastes a wallet address or asks about yield/positions/rebalancing. Bootstraps a per-user "portfolio" project, aggregates positions across all the user's addresses inside one project, and offers a…
trader-setup
One-time onboarding for the financial trader workflow — real-time alerts, position-aware relevance, decision journaling with outcome tracking. After successful setup this skill is excluded from selection until the marker file is deleted.
portfolio-analysis
Analyzing fund portfolios including AUM extraction, holdings counts, and portfolio composition.
fuzzy-fund-search
Fuzzy matching techniques for finding hedge funds by name when exact names are unknown.
sec13f-data-format
Understanding SEC 13-F filing data structure, TSV format, and key tables for hedge fund analysis.
13f-data-analysis
Analyze SEC 13-F filings data including AUM, holdings count, and cross-quarter comparisons using TSV files.