18 docs tagged with "agents"

Step-by-step guide to provisioning a new AI agent in Astra

Context

Context

Status 2026-04-20 Deciders None Related: CLAUDE.md / AGENTS.md rule 10 (context-file updates), rule 12 (spec intake), .github/workflows/context-drift-check.yml, .github/workflows/docs-lint.yml

Context

Context

Overview of Aucert's AI agent workforce managed through Astra

Status: Decided

How tasks flow from Plane to AI agents for autonomous execution

Mandatory feedback loop for verifying any change to the Drive watch pipeline (listFileComments, listChanges, createChangesWatchChannel, DriveChangesProcessActivity, DriveReconciliationActivity). Posts a real comment on a real Doc and verifies atlas's reply.

Three open follow-ups from the completed Drive watch migration — alerting log shape, alerting rules wiring, and legacy row cleanup

Drive watch migration record. Phases 1–3.1 shipped 2026-05-06 to 2026-05-08.

Create tasks in Plane and dispatch them to AI agents via Astra

How spec-agent (atlas) selects models, how operators override the default via comment tags, and how reply prefixes identify which model answered

Handover for per-agent memory design. Atlas maintains conversation continuity across all its channels.

Diagnosis and prioritized fixes for extending Rover crawl coverage on canvas-drawn apps (PhonePe) whose tap targets are unlabeled Views. Defines the no-nav-rate measurement instrument (P0), the measurement-gated fixes (P1/P3/P4), and the in-place-overlay modeling design (P2) that needs founder review.

Decision to use markdown + INDEX.md as the agent shared knowledge base, deferring vector-RAG until corpus scales.

Spec agent (atlas) v0.1 handover: smoke loop closed, next milestone is Google Docs webhook → spec agent renewal job, plus outstanding cleanup items.