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 UnboundCompute/security-agent-skills --skill hunting-formula-injection-in-exportsgit clone --depth 1 https://github.com/UnboundCompute/security-agent-skillsWrote 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/unboundcompute/security-agent-skills/hunting-formula-injection-in-exports)<a href="https://agentmods.dev/skills/unboundcompute/security-agent-skills/hunting-formula-injection-in-exports"><img src="https://agentmods.dev/badge/skills/unboundcompute/security-agent-skills/hunting-formula-injection-in-exports/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/unboundcompute/security-agent-skills/hunting-formula-injection-in-exports"><img src="https://agentmods.dev/badge/skills/unboundcompute/security-agent-skills/hunting-formula-injection-in-exports.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 | $0.00192 | $0.02128 |
| Opus 5 | $0.00096 | $0.01064 |
| Sonnet 5 | $0.00038 | $0.00426 |
| Haiku 4.5 | $0.00019 | $0.00213 |
Grade A, and why
hunting-formula-injection-in-exports 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 7d 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 — 140 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Hunting formula injection in exports: when a cell runs code on open
An application that lets users store text and later offers an export to a spreadsheet format has quietly built a delivery channel. Spreadsheet programs decide whether a cell is data or a formula by its first character: a leading equals, plus, minus, or at sign marks a formula, and so does a leading tab or carriage return that the program strips before parsing. So a value a user typed into a profile field, a comment, or a device name, stored verbatim and written into a downloaded report, becomes live code when an administrator opens that report. The formula can pull other cells out to an attacker URL through a web or hyperlink function, invoke a legacy dynamic-data exchange command that launches a program, or display misleading content the recipient trusts. The bug lives in the export writer, not the input page, and the victim is the person who opens the file. You find it by tracing stored untrusted fields into every export and checking whether dangerous leading characters are neutralized.
When to use
- The application exports user-controlled data to CSV, TSV, or a spreadsheet workbook format.
- A stored field (name, comment, description, tag) is written into a downloaded report opened by staff.
- An admin, finance, or analytics workflow opens exports of user-submitted content in a spreadsheet.
Scope check
Test formula injection only against applications you own or are authorized to assess, and open confirming exports only in an isolated spreadsheet environment with no network egress and no sensitive cells, because a live formula can call out or launch a program on the machine that opens it. Never plant a payload where a real recipient would open it. If you can't name the authorization, stop.
The loop
- Establish that the export writes cells without neutralization first. Locate every code path that produces a CSV, TSV, or workbook export and read how it writes each cell. This is the false-positive killer: if the writer prefixes at-risk cells with a safe character, wraps values so the leading token is inert, forces the cell type to text, or strips leading formula characters, a stored value cannot become a formula and there is no bug. Name the exports that write stored values verbatim.
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.
- 7d ago First seen · 140 lines · 192 tokens per session scan A fe1ac0bdcacf
hunting-formula-injection-in-exports is a skill published in the GitHub repository UnboundCompute/security-agent-skills (5 stars, last pushed 4d ago), licensed MIT. It adds 192 tokens to every session and 2,128 once invoked, about $0.0010 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 2026-09-05.
Other skills, from other repositories
gauntlex:report
Render a stored GAUNTLEX Resilience Report as Markdown, HTML, or structured JSON. Can attach to a GitHub PR comment.
xlsx
Comprehensive spreadsheet creation, editing, and analysis with support for formulas, formatting, data analysis, and visualization. When Claude needs to work with spreadsheets (.xlsx, .xlsm, .csv, .tsv, etc) for: (1) Creating new spreadsheets with formulas and formatting, (2) Reading or analyzing data, (3) Modify…
prodcheck-review
Review this codebase against the prodcheck pre-production checklists — security, performance, scale, integrations and post-launch readiness. Use when asked to check whether a project is ready to ship, to audit an area before launch, or to work through a specific checklist. Produces evidence with file:line citations…
analyzing-malicious-pdf-with-peepdf
Perform static analysis of malicious PDF documents using peepdf, pdfid, and pdf-parser to extract embedded JavaScript, shellcode, and suspicious objects.
analyzing-pdf-malware-with-pdfid
Analyzes malicious PDF files using PDFiD, pdf-parser, and peepdf to identify embedded JavaScript, shellcode, exploits, and suspicious objects without opening the document. Determines the attack vector and extracts embedded payloads for further analysis. Activates for requests involving PDF malware analysis, malicious…
analyzing-macro-malware-in-office-documents
Analyzes malicious VBA macros embedded in Microsoft Office documents (Word, Excel, PowerPoint) to identify download cradles, payload execution, persistence mechanisms, and anti-analysis techniques. Uses olevba, oledump, and VBA deobfuscation to extract the attack chain. Activates for requests involving Office macro…