Zero-Tolerance for Failure

When to Use

• Establish failure-prevention culture and operating norms for mission-critical systems
• Apply HRO principles (preoccupation with failure, reluctance to simplify, sensitivity to operations, commitment to resilience, deference to expertise)
• Design defense-in-depth, fail-safe, and fail-closed controls for software, infra, and OT
• Define verification gates, independent checks, and release hold criteria
• Architect redundancy, isolation, and graceful degradation with explicit failure modes
• Facilitate pre-mortems, FMEA, and risk registers before high-stakes change
• Author stop-the-line policy and escalation when quality or safety signals are ambiguous
• Select and track prevention metrics (defect escape, near-miss, repeat incidents, gate bypass)
• Coach leadership behaviors that counter normalization of deviance and blame theater
• Brief engineering and ops on zero-defect aspiration vs error budgets for the right domain

When NOT to Use

• Own SLI/SLO definitions, error-budget policy, and burn-rate alerting → site-reliability-engineer
• Run live incident war room, SEV classification, and status communications → incident-management-engineer
• Design backup/immutability, RTO/RPO recovery, and ransomware restore architecture → cyber-resilience-engineer
• Enforce CI compile/lint/test gates without broader prevention program → build-validator
• Facilitate sprint ceremonies, backlog grooming, or team agile transformation → agile coaching skills
• Issue HR warnings, performance plans, or legal disciplinary guidance → escalate to HR/legal
• Own classified ATO/accreditation packages without operational excellence deliverables → classified-cyber-security-senior-manager (pair for cleared context)
• Produce ADRs and integration patterns without failure-prevention lens → senior-system-architecture (pair for architecture)

Related skills

Core Workflows

1. Scope, limits, and charter

Clarify what “zero tolerance” means in context—aspiration, gates, and metrics—without perfectionism traps.

See references/zerotolerancescopeandlimits.md.

2. HRO principles and operating mindset

Embed high-reliability organization behaviors in teams that face rare, catastrophic failure.

See references/highreliabilityorganization_principles.md.

3. Prevention, verification, and gates

Layer independent checks, hold points, and evidence before irreversible change.

See references/preventionverificationand_gates.md.

4. Redundancy, degradation, and fail-safe design

Specify failure modes, safe defaults, and degraded operation—not only happy path.

See references/redundancydegradationandfailsafe.md.

5. Pre-mortem, FMEA, and risk registers

Surface latent failures before launch; maintain living risk and mitigation traceability.

See references/premortemfmeaandrisk_registers.md.

6. Leadership, culture, and metrics

Measure prevention; reinforce stop-the-line; reduce normalization of deviance.

See references/leadershipcultureand_metrics.md.

Outputs

• Failure-prevention charter — scope, principles, RACI, interfaces with SRE/IR/QA
• Gate catalog — hold points, owners, evidence required, bypass rules and audit trail
• Design review pack — fail-safe/fail-closed decisions, degradation modes, verification plan
• FMEA / pre-mortem record — failure modes, causes, controls, residual risk, owners
• Stop-the-line policy — triggers, authority, duration, comms, and restart criteria
• Metrics dashboard brief — defect escape, near-miss, repeat incidents, gate effectiveness
• Leadership playbook — behaviors, rituals, and anti-patterns (learning vs blame)

Principles

• Prevent over punish — optimize systems and norms; avoid blame theater and hidden workarounds
• Fail closed by default — ambiguous safety or auth states deny; document explicit fail-open exceptions
• Independent verification — separation between build, check, and approve for high-criticality change
• Deference to expertise — elevate domain experts at the boundary; leaders ask, not override silently
• Measure escapes and near-misses — lagging severity alone rewards luck; track what almost failed
• Stop-the-line is a gift — halting bad change is success; normalize escalation without career penalty
• Pair with peers — reliability math (SRE), incidents (IR program), recovery (resilience), builds (CI)