Agent rollout decision record

GitHub Copilot coding agent

Conditional pilot Assign the coding agent only to selected repos and non-critical issues; require protected branches, CODEOWNERS/human approval, and audit review.

Check this agent for your environment Compare alternatives Methodology

Verdict

Conditional pilot

GitHub-native issue-to-PR workflow is pilotable only with repository installation scope, branch protection, and PR approval rules.

Decision depth: AI-assisted launch evidence record

Evidence audit: AI-assisted evidence audit

Permission footprint

Repository readnot evidenced

File editsnot evidenced

Shell commandsexplicitly evidenced

Pull requestsnot evidenced

Issue/PR commentsnot evidenced

Network/API callsnot evidenced

Secrets/env varsinferred

Sandboxingnot evidenced

Permission controlsnot evidenced

Audit loggingnot evidenced

Enterprise adminnot evidenced

Branch isolationnot evidenced

Human approvalnot evidenced

Production deploymentnot evidenced

Power level

Level 3: executes actions

Highest-risk evidenced power: pull-request creation

Controls separated

Required before pilot — organization must implement

least-privilege permissions, protected branches, mandatory human review, audit logs

Vendor-native control evidenced

branch isolation, security docs, admin console

Vendor-native control not evidenced

native sandbox, permission system, audit logs, SSO, data-retention settings, approval workflow

Pilot plan

Task class: Pull-request and code-review agent

Environment: non-production repository

Maximum access: pull-request creation

Reviewer: repo administrator and code owner

Success criteria: agent can comment or propose PR changes without bypassing branch policy

Stop conditions: organization-wide install or write permission is broader than approved

Prohibited: admin repo permissions, merge without human approval, organization-wide install without review

Blocking issues and unknowns

Blockers: reads repository is not explicitly evidenced in the current source set.

Unknowns: Organization-wide installation scope must be validated in each GitHub org.

Limitations: GitHub permissions determine actual blast radius.

Claim-level evidence

runs commands

GitHub Copilot coding agent: runs commands

Status: explicitly-supported · Reviewer: AI-assisted evidence audit · Confidence: 86/100

Source: About GitHub Copilot cloud agent (repo)

Excerpt: Hooks allow you to execute custom shell commands at key points during agent execution, enabling you to add validation, logging, security scanning, or workflow automation.

Limitations: Official-source excerpt is claim-specific; tenant/workspace configuration can still change rollout risk.

accesses secrets or environment variables

GitHub Copilot coding agent: accesses secrets or environment variables

Status: reasonably-inferred · Reviewer: AI-assisted evidence audit · Confidence: 62/100

Source: About GitHub Copilot cloud agent (repo)

Excerpt: Configure secrets and variables

Limitations: This evidence supports an inference; do not display as simply supported.