GitHub · stars history
decionis/agent-safe-pipeline
SteadyVane has taken 76 readings of decionis/agent-safe-pipeline since Aug 14, 2026: 241 stars at the first reading, 530 at the latest (+289). Its biggest one-day gain was +129 on Aug 14, 2026.
241
first seen stars
535
peak stars
530
latest stars
76
observations
Day by day
| Day (UTC) | stars | Change |
|---|---|---|
| Aug 26, 2026 | 530 | -3 |
| Aug 25, 2026 | 533 | -1 |
| Aug 24, 2026 | 534 | +1 |
| Aug 23, 2026 | 533 | 0 |
| Aug 22, 2026 | 533 | 0 |
| Aug 21, 2026 | 533 | 0 |
| Aug 20, 2026 | 533 | -1 |
| Aug 19, 2026 | 534 | +2 |
| Aug 18, 2026 | 532 | +17 |
| Aug 17, 2026 | 515 | +31 |
| Aug 16, 2026 | 484 | +44 |
| Aug 15, 2026 | 440 | +70 |
| Aug 14, 2026 | 370 | +129 |
| Description | Reference architecture for AI agents that propose actions but cannot authorize them — immutable intent capture, an independent Decionis policy verdict (ALLOW/ESCALATE/BLOCK), verified human approval, and a SafeExecutor that consumes a single-use intent-bound grant. |
|---|---|
| Language | TypeScript |
| Topics | agentic-ai, ai-agent-permissions, ai-agents, ai-governance, ai-safety, authorization, decionis, execution-authority |