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 seonglae/openportfolio --skill openportfolio-recordgit clone --depth 1 https://github.com/seonglae/openportfolioWrote 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/seonglae/openportfolio/openportfolio-record)<a href="https://agentmods.dev/skills/seonglae/openportfolio/openportfolio-record"><img src="https://agentmods.dev/badge/skills/seonglae/openportfolio/openportfolio-record/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/seonglae/openportfolio/openportfolio-record"><img src="https://agentmods.dev/badge/skills/seonglae/openportfolio/openportfolio-record.svg" alt="Reviewed on agentmods" width="80" height="20"></a>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 | $0.00071 | $0.00652 |
| Opus 5.5 | $0.00028 | $0.00261 |
| Sonnet 5.5 | $0.00014 | $0.00130 |
| Haiku 4.5 | $0.00007 | $0.00065 |
Grade A, and why
openportfolio-record 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 yesterday.
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 — 33 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Checking the record
Needs the openportfolio MCP server. See the openportfolio-book skill for setup.
Reading it
calibration returns the mean Brier score, the expected calibration error, and the reliability buckets behind them. Optional windowDays, author and buckets.
- Brier is the mean of
(probability - outcome)^2with outcome 1 or 0. Lower is better. - 0.25 is what saying 50% to everything earns; beating it alone does not establish skill. Read paired baseline scores and their shared count. Open, void and legacy unaudited settlements are reported separately. No confidence interval is claimed because event dependence is unknown.
- Read the buckets, not only the mean. Each bucket compares what was said with how often it happened. Observed below said at the top end is overconfidence, the usual failure. A bucket with three calls in it says very little, so report the count beside the rate.
- Filter by
authorto inspect a forecaster. Means on different event sets are not a fair head-to-head comparison.
list_forecasts (optional status of open, resolved or void, author, limit) lists the calls themselves, newest first.
Settling what is due
due_forecasts lists open calls whose horizon has passed. requiresHistoricalEvidence false means no worker will settle the row.
- A row with a supported resolutionTarget can be settled by the worker from historical evidence under the policy frozen at registration. A symbol alone is insufficient. For
settle_forecast, observedValue requires evidence matching the assetId, symbol, assetClass, venue, currency and registered time policy, with a source. Never use a current price for an old horizon or widen the allowed age after seeing prices. - Legacy symbol-only calls and stored unsupported targets require a documented manual outcome and source note. Calls with a supported target cannot bypass it with a manual outcome. Both paths reject early settlement; do not pass outcome and observedValue together.
- A row with a null symbol is prose and needs reading. Establish whether the criterion was met, then
settle_forecastwith an explicitoutcomeof true or false and anotesaying how you know. - If the criterion became unresolvable (the event was cancelled or the wording turned out ambiguous), use
void_forecastwith areason. A void drops out of the mean. Never settle a call you cannot establish: a guess scored as an outcome corrupts the record it is meant to test.
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.
- yesterday Changed · +3 lines 2598de639545
- 16d ago First seen · 30 lines · 71 tokens per session scan A f35e553ee243
openportfolio-record is a skill published in the GitHub repository seonglae/openportfolio (3 stars, last pushed 2d ago), licensed Apache-2.0. It adds 71 tokens to every session and 652 once invoked, about $0.0003 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 2026-09-30.
Other skills, from other repositories
research-stocks
Read current stock research, technical backtests, saved analyses or saved backtests. Use for a named ticker or watchlist research history; preserve source dates, model limitations and risk factors.
track-tariff-trends
Retrieve a country's average applied tariff over time — the MFN rate on all products, or the rate it applies on goods from one partner and/or in one product group — plus the optional US effective tariff rate. Use when the user asks how a country's tariffs have changed, or what it charges another country on average.…
get-prediction-markets
Retrieve active prediction-market contracts (Polymarket) with live yes-price probabilities, volume, and close dates, filterable by category or keyword. Use when the user asks what the market odds are on a geopolitical, economic, or election outcome.
tvchart
Drive a live TradingView Lightweight Charts widget end-to-end via PyWry MCP tools — symbol, interval, indicators, markers, price lines, layouts, state.
tbank-messenger
A chat connection for T-Bank, a Russian bank, that lets an agent read conversations, download attachments, and contact bank support. It works with chats inside the bank app, not SMS codes or push notifications.
aifi-daily-report
A tool for producing a Chinese end-of-day report on United States stock markets as a self-contained HTML file. HTML is the standard format used to display web pages.