governance/start-agent-session

Create deterministic agent session records with validated paths and optional U.Work logging.

Updated Jan 11, 2026
One-click install
npx skills add https://github.com/venikman/fpf-foundry --skill governance-start-agent-session
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: governance/start-agent-session
Source: https://github.com/venikman/fpf-foundry/tree/main/design/skills/governance/start-agent-session
Command: npx skills add https://github.com/venikman/fpf-foundry --skill governance-start-agent-session

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

This skill provides a safe, auditable way to start a new agent session record within a bounded context, establishing a deterministic anchor for governance handoffs and activity tracking.

Core Features & Use Cases

  • Create runtime/contexts/<context>/sessions/<session_id>.session.md with status set to active and started_at timestamp.
  • Validate inputs to ensure safe path segments and prevent collisions or overwrites.
  • Optionally emit a U.Work log via telemetry/log-work when available to capture initiation details and agent metadata.

Quick Start

To start a session, supply context, session_id, and a descriptive title. Example: context "AgentWorkflow", session_id "AW-0001", title "Agent workflow bootstrap", and optional capability_ref "runtime/contexts/AgentWorkflow/capabilities/default.capability.yaml".

Frequently Asked Questions about governance/start-agent-session

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I create a deterministic agent session record for audit logging?

To create a deterministic agent session record, you supply a bounded context, session_id, and title. The process writes a markdown file to runtime/contexts/<context>/sessions/, sets status to active, and adds a started_at timestamp for governance tracking.

What is a bounded context in agent session governance and when do I need it?

A bounded context isolates agent activity within a specific domain boundary for governance and audit tracking. You need it when establishing deterministic session anchors to ensure safe path segments, prevent file collisions, and maintain accurate telemetry records.

Can I log agent session telemetry automatically when starting a session?

Yes, you can optionally log agent session telemetry by emitting a U.Work log entry via telemetry/log-work when the module is available, capturing initiation details and agent metadata alongside the session record creation.

How do I anchor governance decisions and handoffs within an agent session?

You anchor governance decisions and handoffs by supplying an optional capability reference and a decisions list when starting the session. This links the session record to specific governance rules and documented handoff points.

Does starting an agent session require any external dependencies or components?

No external dependencies are required to start an agent session. The core functionality validates inputs, ensures safe path segments, and creates directories as needed. Telemetry logging is optional and only activates if the telemetry/log-work module is present.

What are the limitations when creating agent session records for audit trails?

Limitations include strict input validation for safe path segments to prevent collisions or overwrites, meaning malformed context or session_id values will be rejected. Additionally, automatic U.Work telemetry logging only functions if the telemetry/log-work module is available.