Roadmap

gh-aw-fleet is pre-1.0. The roadmap is biased toward operator ergonomics, cost visibility, and supply-chain safety while preserving the thin-orchestrator boundary: delegate to gh aw, gh, and git instead of re-implementing them.

See the full ROADMAP.md for the live issue list and status notes.

Release composition

Each feature-bearing release should include at least one cost-visibility item and at least one security item. Cost matters because usage-based Copilot billing made AI-credit spend a continuous operator concern. Security matters because fleet is the layer that knows every workflow pin across every managed repo.

Single-fix hotfix releases are exempt.

Immediate focus

The current release focus pairs one cost item with one security item:

• Read-only over-budget highlighting in the consumption rollup (--budget).
• Promotion of high-severity Layer 1 security findings from advisory to blocking under an opt-in --strict flag.

Near-term FinOps

Planned cost-visibility work includes:

• gh-aw-fleet forecast: aggregate projected AIC before spend happens.
• Pagination for Actions workflow discovery in the logs consumption source.
• Better cap-hit diagnostic hints when max AI credits or max turns are exceeded.

Security

Security work continues through the scanner pipeline:

• Layer 1 scanner coverage for secrets, compiled workflow hazards, and fleet-structural rules.
• --strict promotion for high findings.
• Future deeper scans for prompt-injection signatures and risky trigger patterns.

Documentation and operator quality of life

The roadmap also tracks install friction, dry-run examples, troubleshooting hints, and packaging as a gh extension. Documentation work should stay close to real operator workflows: installation, configuration, reconcile, troubleshooting, and cost review.