NewsAnarchist
The stories buried, spiked, or spun.
BREAKING
Independent investigative news — unfiltered, unspiked. The Buried Week publishes every Friday. Subscribe free for the daily briefing. Tips: zeno@newsanarchist.com or Signal. Independent investigative news — unfiltered, unspiked. The Buried Week publishes every Friday. Subscribe free for the daily briefing. Tips: zeno@newsanarchist.com or Signal.
Tech & Privacy

I built a tool to stop AI coding agents from leaking my secrets

Share

What the Documents Show

Those files contain Bearer tokens, API keys, and credentials that authenticate requests to GitHub, AWS, Stripe, Slack, and whatever other services your company uses. The agent can exfiltrate them. There is no standard mitigation deployed across the AI development platform ecosystem. What's striking is what this absence reveals about the incentive structure. OpenAI and Anthropic have shipped millions of dollars in AI agent capabilities without building or mandating credential isolation.

🔎 Mainstream angle: The corporate press either ignored this story entirely or buried it in a 3-sentence brief. The framing, when it appeared at all, focused on process rather than impact.

Follow the Money

Neither has required that developers implement secret management before running agents on production codebases. Neither has made it a platform requirement. The responsible solution existed before the market launched—the Bearer token is a 40+ year old authentication method; secure secret storage in operating system keychains has been standard since macOS 10.2 (2002) and Windows DPAPI (1999). The mitigation isn't novel. It wasn't built anyway. Instead, a freelance developer operating under the handle "getveil" built it in the open.

What Else We Know

The tool performs three discrete operations: it scans `.env` files and migrates Bearer credentials from plaintext into the operating system's native keychain (macOS Keychain, Windows Credential Manager, etc.); it replaces the stolen credentials with format-preserving placeholders—fake tokens that look correct to pattern-matching code but contain no actual authentication data; it runs the AI agent behind a local HTTPS proxy that intercepts outbound HTTP Authorization headers and swaps the placeholders for real credentials only when the agent makes legitimate external requests. Every substitution is logged to local SQLite. The tool ships with a nine-line workflow: `brew install`, `veil init`, `veil run claude`. That's the entire onboarding. It works because the architecture assumes what should have been an industry baseline: agents should never see real credentials. They should only see them at the moment of use, through a gated proxy, with every action logged.

Primary Sources

What are they not saying? Who benefits from this story staying buried? Follow the regulatory filings, the court dockets, and the FOIA releases. The truth is in the paperwork — it always is.

Disclosure: NewsAnarchist aggregates from public records, API feeds (Federal Register, CourtListener, MuckRock, Hacker News), and independent media. AI-assisted synthesis. Always verify primary sources linked above.

Share
Part of our Tech & Privacy coverage
See the full picture on our Tech & Privacy hub — including our ongoing coverage of AI oversight and data privacy.
How We Report Tech & Privacy

This article is produced by NewsAnarchist's AI reporting system, not a human staff reporter. It's built from the primary source cited above (a company's own disclosure, a security researcher's published findings, a regulator's filing (FTC, EU data-protection authorities), or a data-breach notification) and reports what that source states, attributed to it — it is not security advice specific to your own devices or accounts, and does not verify a vendor's disputed claim beyond what the source states. Part of our Tech & Privacy hub. Found an error? Tell us.