Skip to content

v0.1.0

Latest
Latest

Choose a tag to compare

View all tags
@jsilets jsilets released this 11 Jul 15:25
· 51 commits to main since this release
v0.1.0
ac1ad76
This commit was created on GitHub.com and signed with GitHub’s verified signature.
GPG key ID: B5690EEEBB952194
Verified
Learn about vigilant mode.

First tagged release. The connect → accept → message loop has run against live LinkedIn on a single supervised account with the safety gate enforced server-side.

Added

  • Control plane over MCP: campaign planning, sourcing and lead-list tools, sequence definition, and an approval flow for supervised autonomy.
  • Per-account runner driving a real headful browser (session, safety pre-flight, executor, reply detector) with an append-only audit log.
  • Server-side SafetyGate: daily caps, working hours and days, account state machine, operator pause, and approval-time enforcement.
  • Driven mode (an external agent as the brain over MCP) and autonomous mode (an internal LLM loop), both gated the same way server-side.
  • Campaign-visibility web dashboard: funnel, per-lead table, in-UI approvals.
  • Single-image Docker deployment (Xvfb + headful Chromium) and a Railway config, with LOA_EXECUTOR=fake as the default for safe exploration.

Deploy defaults to LOA_EXECUTOR=fake; see infra/RAILWAY.md to switch a single supervised account to real sending.

Assets 2
Loading