Problem statement
100% of simulated learners (across all levels, personalities, and tools) fail at 08-run-your-workflow in every one of 1000 Monte Carlo runs. The recorded failure category is exclusively workflow-not-compiled: the step's pre-flight section (## Pre-flight check) re-checks the same lock-file and billing-permission state already verified in Step 7 and Step 7d, but offers no way for the learner to tell which of the two conditions (stale lock file vs. mismatched billing method) actually caused their specific run to fail. The only remediation offered is "return to Step 7d," with no differential diagnosis.
Proposed change
Add a short troubleshooting decision table/flow to workshop/08-run-your-workflow.md (in or near the "Pre-flight check" section) that helps the learner read the actual run-failure error text and map it to one of two concrete fixes:
401 Unauthorized / permission-denied on the Copilot API call → billing method mismatch → link to the specific Path A/Path B fix in side-quest-07d-billing-paths.md- Workflow missing from the Actions tab entirely, or run fails immediately with no logs → stale/missing
.lock.yml → recompile-and-push command block
Failure mode classification
Access barrier (environment/tooling/auth state, not conceptual difficulty).
Quantitative guardrail
Current 08-run-your-workflow.md scores: overall_score: 5.67/10, cognitive_load: 9.1, active_learning: 3.0, checkpoint_quality: 0.0, scaffolding: 5.0, style_compliance: 10.0, readability: 10.0. Learning KPI index for this step: (2.0×3.0 + 2.0×0.0 + 1.5×5.0) / 5.5 = 2.45/10.
Adding a differential-diagnosis block increases scaffolding (more support for recovering from a stuck state) and active_learning (learner reads their own error and self-diagnoses rather than blindly retrying), without adding new required tool installs or removing any existing verification step — so overall_score and the learning KPI index should hold flat or rise, not fall.
Acceptance criteria
Suggested owner
copilot coding agent
Related to #3195
Generated by 🔬 Workshop Student Simulator · copilot · auto · 186.5 AIC · ⌖ 20.1 AIC · ⊞ 15.2K · ◷
Problem statement
100% of simulated learners (across all levels, personalities, and tools) fail at
08-run-your-workflowin every one of 1000 Monte Carlo runs. The recorded failure category is exclusivelyworkflow-not-compiled: the step's pre-flight section (## Pre-flight check) re-checks the same lock-file and billing-permission state already verified in Step 7 and Step 7d, but offers no way for the learner to tell which of the two conditions (stale lock file vs. mismatched billing method) actually caused their specific run to fail. The only remediation offered is "return to Step 7d," with no differential diagnosis.Proposed change
Add a short troubleshooting decision table/flow to
workshop/08-run-your-workflow.md(in or near the "Pre-flight check" section) that helps the learner read the actual run-failure error text and map it to one of two concrete fixes:401 Unauthorized/ permission-denied on the Copilot API call → billing method mismatch → link to the specific Path A/Path B fix inside-quest-07d-billing-paths.md.lock.yml→ recompile-and-push command blockFailure mode classification
Access barrier (environment/tooling/auth state, not conceptual difficulty).
Quantitative guardrail
Current
08-run-your-workflow.mdscores:overall_score: 5.67/10,cognitive_load: 9.1,active_learning: 3.0,checkpoint_quality: 0.0,scaffolding: 5.0,style_compliance: 10.0,readability: 10.0. Learning KPI index for this step:(2.0×3.0 + 2.0×0.0 + 1.5×5.0) / 5.5 = 2.45/10.Adding a differential-diagnosis block increases
scaffolding(more support for recovering from a stuck state) andactive_learning(learner reads their own error and self-diagnoses rather than blindly retrying), without adding new required tool installs or removing any existing verification step — sooverall_scoreand the learning KPI index should hold flat or rise, not fall.Acceptance criteria
overall_scorefor08-run-your-workflow.mdstays flat or improves after the change(2.0 × active_learning + 2.0 × checkpoint_quality + 1.5 × scaffolding) / 5.5stays flat or improvesSuggested owner
copilot coding agentRelated to #3195