Summary
The local HTTP server started by engram server (binding 127.0.0.1:7337 by default) was exposed to any browser origin with no authentication unless ENGRAM_API_TOKEN was explicitly set. Combined with Access-Control-Allow-Origin: * on every response and a body parser that did not require Content-Type: application/json, this allowed a malicious web page the developer visited to:
- Exfiltrate the local knowledge graph via
GET /query and GET /stats (function names, file layout, recorded decisions/mistakes).
- Inject persistent prompt-injection payloads via
POST /learn, which wrote mistake/decision nodes that were later surfaced as system-reminders to the user's AI coding agent on every future session and file edit.
Severity: High — confidentiality + persistent indirect prompt injection against the user's coding agent.
Affected versions
engramx >= 1.0.0, < 2.0.2 — any version that shipped the HTTP server.
Patched in
engramx@2.0.2
Workarounds (if you cannot upgrade)
- Do not run
engram server or engram ui.
- If developers must, set
ENGRAM_API_TOKEN to a long random value and terminate the server before browsing the web.
Remediation (applied in 2.0.2)
- Fail-closed auth on every non-public route — Bearer header or HttpOnly cookie, constant-time comparison, 256-bit auto-generated token at
~/.engram/http-server.token (0600).
- Wildcard CORS removed entirely; default is no CORS headers. Opt-in allowlist via
ENGRAM_ALLOWED_ORIGINS.
- Host + Origin validation — rejects DNS rebinding and Host spoofing.
Content-Type: application/json enforced on mutations — blocks the text/plain CSRF vector.
/ui?token= bootstrap with Sec-Fetch-Site gate — prevents cross-origin oracle probing.
Credit
Discovered and responsibly disclosed by @gabiudrescu in engram issue #7.
References
Summary
The local HTTP server started by
engram server(binding127.0.0.1:7337by default) was exposed to any browser origin with no authentication unlessENGRAM_API_TOKENwas explicitly set. Combined withAccess-Control-Allow-Origin: *on every response and a body parser that did not requireContent-Type: application/json, this allowed a malicious web page the developer visited to:GET /queryandGET /stats(function names, file layout, recorded decisions/mistakes).POST /learn, which wrotemistake/decisionnodes that were later surfaced as system-reminders to the user's AI coding agent on every future session and file edit.Severity: High — confidentiality + persistent indirect prompt injection against the user's coding agent.
Affected versions
engramx>= 1.0.0, < 2.0.2 — any version that shipped the HTTP server.Patched in
engramx@2.0.2Workarounds (if you cannot upgrade)
engram serverorengram ui.ENGRAM_API_TOKENto a long random value and terminate the server before browsing the web.Remediation (applied in 2.0.2)
~/.engram/http-server.token(0600).ENGRAM_ALLOWED_ORIGINS.Content-Type: application/jsonenforced on mutations — blocks the text/plain CSRF vector./ui?token=bootstrap withSec-Fetch-Sitegate — prevents cross-origin oracle probing.Credit
Discovered and responsibly disclosed by @gabiudrescu in engram issue #7.
References