mawkone 7a601b57b8 fix: rotate Gitea token and prevent empty assistant messages in history ...

- Empty message fix: skip pushing assistant msg to history when both
  content and tool_calls are absent (GLM-5 mid-reasoning token exhaustion).
  Also filter preexisting empty assistant messages from returned history.
- System prompt now correctly injects knowledgeContext from opts into the
  Tier-B system message (was missing in the loop's buildMessages).
- GITEA_API_TOKEN updated externally in Coolify (old token was invalid).

Made-with: Cursor

2026-02-27 20:17:34 -08:00

dist

fix: decode GCP_SA_KEY_BASE64 for Vertex AI auth (avoids Dockerfile injection issues)

2026-02-27 20:00:21 -08:00

node_modules

init: vibn-agent-runner — Gemini autonomous agent backend

2026-02-26 14:50:20 -08:00

src

fix: rotate Gitea token and prevent empty assistant messages in history

2026-02-27 20:17:34 -08:00

.env.example

fix: add all missing source files (llm.ts, updated agent-runner/agents)

2026-02-27 19:33:59 -08:00

Dockerfile

fix: force NODE_ENV=development during npm ci so tsc is installed

2026-02-27 19:30:40 -08:00

package-lock.json

fix: replace gcloud shell-out with google-auth-library for Vertex AI tokens

2026-02-27 19:38:07 -08:00

package.json

fix: replace gcloud shell-out with google-auth-library for Vertex AI tokens

2026-02-27 19:38:07 -08:00

PROJECT.md

feat: Add GET /api/status endpoint with job stats and update PROJECT.md

2026-02-26 23:30:15 +00:00

tsconfig.json

init: vibn-agent-runner — Gemini autonomous agent backend

2026-02-26 14:50:20 -08:00

Description

No description provided

64 MiB

Languages

HTML 45.8%

TypeScript 30%

JavaScript 10.9%

Python 7.6%

CSS 4.7%

Other 0.9%