Risk-based autonomy means the amount of work an agent may finish without a person depends on how risky and reversible the change is, and grows only as verification and rollback prove trustworthy. All three sources that discuss it keep production behind a human gate.

Sizing by path (GH-600)

RiskExample pathsDesign
Lowdocs/, formattingAuto-merge after required checks
Mediumsrc/, dependency bumpsPR, checks, at least one review
Highinfra/, .github/workflows/CODEOWNERS, multiple reviews, stricter rulesets
CriticalProduction deploys, settings, secretsEnvironment approval: the agent prepares but does not execute

As described in module 2.1 A GitHub Actions environment with required reviewers pauses any job targeting it until a person approves.1

Widening in steps (AI-native SDLC playbook)

  1. Start with read-only work such as build-failure triage and changelog drafts.
  2. Let agents propose changes only through pull requests and branch protection.
  3. Run jobs in sandboxes with short-lived, scoped credentials.
  4. Expose deployment and rollback as allowlisted tools per environment.
  5. Allow more autonomy in development than in production.
  6. Keep the production gate human and rehearse rollback.

2

Rolling out Claude Code Auto Mode

The Auto Mode guidance follows the same shape: start narrow, keep explicit deny and ask rules, watch what is denied, widen gradually, and keep human review for production infrastructure.3

Reliability assumptions

GH-600 adds that agent workflows should assume failure: bounded retries for transient check failures, escalation to a person after a check fails twice (with what failed, what was tried, and a suggested next step), and rollback readiness for high-risk changes.1

Footnotes

  1. GitHub Certified: Agentic AI Developer (study notes) 2 3

  2. The AI-Native SDLC Playbook

  3. How Claude Code Auto Mode Works