| name | sop-dogfooding-quality-detection |
| description | SOP for detecting quality regressions during dogfooding runs and turning them into actionable fixes. |
| allowed-tools | Read, Write, Edit, Bash, Glob, Grep, Task, TodoWrite |
| model | sonnet |
| x-version | 3.2.0 |
| x-category | quality |
| x-vcl-compliance | v3.1.1 |
| x-cognitive-frames | HON, MOR, COM, CLS, EVD, ASP, SPC |
STANDARD OPERATING PROCEDURE
Purpose
Identify quality regressions and latent issues while dogfooding, ensuring findings are evidenced, prioritized, and fed back into improvement loops.
Trigger Conditions
- Positive: active dogfooding sessions, regression sweeps after releases, or monitoring new features for emergent issues.
- Negative: isolated bug triage without self-application or pattern capture.
Guardrails
- Confidence ceiling: Use
Confidence: X.XX (ceiling: TYPE Y.YY)with ceilings {inference/report 0.70, research 0.85, observation/definition 0.95}. - Evidence-first: Record file:line, logs, metrics, or reproduction steps for each detected issue.
- Structure-first: Update examples/tests to reflect newly detected regressions and their fixes.
- Prioritization: Tag severity and blast radius; block release on critical regressions until resolved or waived with rationale.
Execution Phases
- Observation & Capture
- Monitor outputs, logs, and behaviors during dogfooding; collect anomalies.
- Normalize entries with severity, location, and reproduction notes.
- Validation & Classification
- Reproduce findings; distinguish false positives and intentional behavior.
- Map to categories (correctness, performance, UX, security, reliability).
- Remediation & Feedback
- Propose fixes and owners; add tests to prevent recurrence.
- Feed learnings into pattern retrieval and references.
- Confidence & Closure
- Confirm fixes or document waivers; state residual risk and confidence with ceiling.
Output Format
- Log of detected issues with evidence and severity.
- Reproduction steps and validation results.
- Remediation plan and test updates.
- Confidence statement using ceiling syntax.
Validation Checklist
- Evidence captured with location/steps for each issue.
- False positives filtered; categories assigned.
- Fixes/tests identified and owners named.
- Patterns/references updated where applicable.
- Confidence ceiling provided; English-only output.
Confidence: 0.70 (ceiling: inference 0.70) - SOP rewritten per Prompt Architect confidence discipline and Skill Forge structure-first detection loop.