Proposal: Inject secrets into HTTP headers using MITM proxy

[jakub-bochenski]

I think this is a more robust approach than https://github.com/mattolson/agent-sandbox/blob/main/docs/plan/milestones/m13-host-credential-service/milestone.md

[mattolson]

I'm not sure I understand, can you elaborate?

[jakub-bochenski]

E.g. this project implements secret injecton: https://github.com/jingkaihe/matchlock/

[jakub-bochenski]

As per #103 there is a ready secret injection setup for docker compose that could be reused here

E.g.

• https://github.com/VirtusLab/sandcat/blob/5faa9a2/scripts/mitmproxy_addon.py#L121
• https://github.com/VirtusLab/sandcat/blob/5faa9a2dd22e43b3016a13b266aefb7166227a42/README.md#secret-substitution

[mattolson]

I like this idea - I'll update the m13 milestone with the new approach.

[mattolson]

Wrote up a plan here: #123