Description
The copilot-CLI Execute GitHub Copilot CLI 0-turn startup failure is now the single dominant prod-main failure family — 100% of failures in both the Weekly Workflow Analysis (#43770, 7/7) and the Agentic Workflow Audit (#43629, 11/11) came from this one step. It has two chronic clusters: Smoke CI failing 100% on push to main (recurrence #5) and SDK-longrun runs that do 10-24 minutes of real work then fail 0-turn at the CLI step (recurrence #23). Long-run workflows burn 10-24 min of agent work (e.g. Daily Safe Output Integrator: 24m -> 0-turn, 74 AIC wasted) before the failure surfaces.
What to do
Add a fast-fail copilot-CLI startup diagnostic that probes CLI readiness at the start of the agent job and surfaces why the CLI aborts 0-turn, so affected workflows fail in seconds rather than after minutes of work. Emit the abort reason (auth, availability, config) into the step summary. This does not fix the root cause (engine-owner escalation) but stops the wasted action-minutes and gives clear signal.
Expected Impact
Cuts wasted action-minutes/AIC on the fleet's dominant failure family; converts opaque 0-turn reds into actionable diagnostics.
Suggested Agent
New Agent or copilot-engine maintainer (touches the engine wrapper).
Estimated Effort
Medium (1-4 hours)
Data Source
DeepReport Intelligence Briefing 2026-07-06; source discussions #43770, #43629.
Generated by 🔬 Deep Report · 204.6 AIC · ⌖ 17.5 AIC · ⊞ 10.2K · ◷
Description
The copilot-CLI
Execute GitHub Copilot CLI0-turn startup failure is now the single dominant prod-main failure family — 100% of failures in both the Weekly Workflow Analysis (#43770, 7/7) and the Agentic Workflow Audit (#43629, 11/11) came from this one step. It has two chronic clusters: Smoke CI failing 100% on push to main (recurrence #5) and SDK-longrun runs that do 10-24 minutes of real work then fail 0-turn at the CLI step (recurrence #23). Long-run workflows burn 10-24 min of agent work (e.g. Daily Safe Output Integrator: 24m -> 0-turn, 74 AIC wasted) before the failure surfaces.What to do
Add a fast-fail copilot-CLI startup diagnostic that probes CLI readiness at the start of the
agentjob and surfaces why the CLI aborts 0-turn, so affected workflows fail in seconds rather than after minutes of work. Emit the abort reason (auth, availability, config) into the step summary. This does not fix the root cause (engine-owner escalation) but stops the wasted action-minutes and gives clear signal.Expected Impact
Cuts wasted action-minutes/AIC on the fleet's dominant failure family; converts opaque 0-turn reds into actionable diagnostics.
Suggested Agent
New Agent or copilot-engine maintainer (touches the engine wrapper).
Estimated Effort
Medium (1-4 hours)
Data Source
DeepReport Intelligence Briefing 2026-07-06; source discussions #43770, #43629.