Follow-up R3 from the microsoft/AI-Engineering-Coach exploration (docs/superpowers/explorations/2026-06-16-ai-engineering-coach-findings.md, issue #340). Suggested first build of the set — smallest, fully additive, needs no behavioural data.
Idea
The MS Context Health page emits an agentic-readiness checklist + scalar score. We have harness-audit-engine drift detection, but it produces a backward-looking drift report, not a forward-looking readiness view. Add a binary readiness checklist (with a small score) answering "is this workspace set up for agentic work?"
Candidate checklist items (all from existing artifacts)
- Instruction files present and non-stub: CLAUDE.md, HARNESS.md, AGENTS.md.
- MODEL_ROUTING.md present with routing rules.
- Constraints enforced (not all
unverified).
- Reflection loop active (recent
reflections/active/ fragments).
- Observability snapshot fresh (< cadence threshold).
- Orchestrator / agent pipeline present (if applicable).
Home
Extend /harness-audit or /superpowers-status with a readiness section, or a small new subcommand. Reuse harness-audit-engine surface scanning. Add an output-validation checkpoint per the repo convention.
Note
Distinct from drift detection (which asks "has it gone stale?") — this asks "is it ready?". Both are useful; this is the missing forward-looking half.
Follow-up R3 from the microsoft/AI-Engineering-Coach exploration (
docs/superpowers/explorations/2026-06-16-ai-engineering-coach-findings.md, issue #340). Suggested first build of the set — smallest, fully additive, needs no behavioural data.Idea
The MS Context Health page emits an agentic-readiness checklist + scalar score. We have
harness-audit-enginedrift detection, but it produces a backward-looking drift report, not a forward-looking readiness view. Add a binary readiness checklist (with a small score) answering "is this workspace set up for agentic work?"Candidate checklist items (all from existing artifacts)
unverified).reflections/active/fragments).Home
Extend
/harness-auditor/superpowers-statuswith a readiness section, or a small new subcommand. Reuseharness-audit-enginesurface scanning. Add an output-validation checkpoint per the repo convention.Note
Distinct from drift detection (which asks "has it gone stale?") — this asks "is it ready?". Both are useful; this is the missing forward-looking half.