Security · Disclosure

Responsible
disclosure

If you believe you've found a security issue, report it privately so a fix can ship before public discussion. Do not open public GitHub issues for vulnerabilities.

Channel
security@openagent.id
Mode
Coordinated
Public issues
Not for vulns
§ 01
Reporting

What to send

Good reports minimize back-and-forth: a reproduction path, impact, and enough context for triage.

email · security@openagent.id

Channel

Email security@openagent.id. If your report is sensitive, keep details minimal until we confirm a secure follow-up channel.

Reference the repository and file paths if possible. Include the smallest reproduction that demonstrates impact.

goal · Triage-ready

Checklist

01What you expected vs. what happened
02Minimal reproduction steps or proof-of-concept
03Impact assessment (what an attacker gains)
04Affected repository/package + versions (if known)
05Any logs or traces that help reproduce safely
§ 02
Targets

Response targets

These are targets, not guarantees. The goal is to acknowledge quickly, reproduce safely, and coordinate disclosure responsibly.

SeverityInitialTriageFix
CriticalAim: 1 business dayAim: 2 business daysCoordinate disclosure
HighAim: 3 business daysAim: 7 daysCoordinate disclosure
MediumAim: 7 daysAim: 14 daysCoordinate disclosure
LowAim: 14 daysBest effortBest effort
§ 03
Good-faith research

How to help without causing harm

We welcome good-faith security research. Please avoid privacy violations, data destruction, and service disruption.

Minimize harm

Please do

01Use test accounts and non-production data where possible
02Minimize exploitation: demonstrate impact, then stop
03Report privately with reproduction steps
04Coordinate public disclosure timing after a fix exists

Out of scope

Please avoid

01Accessing or exfiltrating data that isn't yours
02Disrupting availability (including volumetric DDoS)
03Social engineering maintainers
04Publicly disclosing before a fix ships
Next

Prefer prevention?

If you're reviewing the security model or threat vectors, start with the threat model and primitive choices.