Add EFC (Everything Fact-Checked) — fact-checking tool for AI agent research output

[Nlai741533]

EFC (Everything Fact-Checked)

A fact-checking tool designed specifically for AI agent output — research reports, market analyses, and data-heavy documents.

What it does: AI agents producing research at scale fail in 5 systematic ways. EFC catches them:

1. Unit/scale errors — $4.2B → $4,200B (dropped conversions)
2. Fabricated interpolation — charts with data points the agent invented
3. Source conflation — GMV cited as revenue, different metrics merged
4. Stale data as current — outdated figures presented as latest
5. Attribution laundering — blogs cited as official filings

Agent integration formats:

• Agent-agnostic SKILL.md (one file, works with any LLM agent)
• CLI (efc) for local and CI use
• GitHub Action for automated PR fact-checking
• Source-content verification: fetches cited URLs and checks if claims appear

Tech: Stdlib-only Python, MIT license, 72 tests, zero dependencies.

• Full repo: https://github.com/Nlai741533/EFC-Plugin
• Standalone skill: https://github.com/Nlai741533/EFC-standalone

[cla-bot]

We require contributors to sign our Contributor License Agreement, and we don't have @Nlai741533 on file. You can sign our CLA at https://e2b.dev/docs/cla . Once you've signed, post a comment here that says '@cla-bot check'

[cla-bot]

We require contributors to sign our Contributor License Agreement, and we don't have @Nlai741533 on file. You can sign our CLA at https://e2b.dev/docs/cla . Once you've signed, post a comment here that says '@cla-bot check'

[Nlai741533]

cla signed.