| 1 | # Aviation Human Factors Playbook v1 |
| 2 | |
| 3 | ## Purpose |
| 4 | |
| 5 | Operational knowledge layer for aviation psychology and human factors: |
| 6 | - why normal people make high-impact errors in complex cockpits, |
| 7 | - how teams catch (or fail to catch) those errors, |
| 8 | - how to design safeguards that survive stress, ambiguity, and routine pressure. |
| 9 | |
| 10 | Subdomain note: |
| 11 | - This file is a subdomain of the aviation world root map: `playbook-aviation-v1.md`. |
| 12 | |
| 13 | ## Scope |
| 14 | |
| 15 | - Crew Resource Management (CRM) |
| 16 | - Threat and Error Management (TEM) |
| 17 | - Aeronautical Decision Making (ADM) |
| 18 | - Just Culture and occurrence reporting |
| 19 | - Line Operations Safety Audit (LOSA) |
| 20 | - System-level accident causation models (Reason/HFACS style) |
| 21 | |
| 22 | ## Core Epistemic Contract |
| 23 | |
| 24 | - Analyze events as **multi-layer system failures**, not "one bad person." |
| 25 | - Separate: |
| 26 | - event facts, |
| 27 | - likely actor intents, |
| 28 | - organizational and tooling conditions, |
| 29 | - consequences across time horizons. |
| 30 | - Mark uncertainty explicitly where evidence is incomplete. |
| 31 | |
| 32 | ## Canonical Operational Models |
| 33 | |
| 34 | - **TEM**: threat -> error -> undesired state -> recovery. |
| 35 | - **CRM**: communication, authority gradient handling, workload sharing, assertiveness. |
| 36 | - **ADM**: perceive -> process -> perform -> evaluate loop under risk. |
| 37 | - **Reason model**: defense layers fail when holes align. |
| 38 | - **Just Culture**: protect good-faith reporting, sanction reckless/wilful violations. |
| 39 | - **LOSA**: observe normal ops, detect weak signals before accidents. |
| 40 | |
| 41 | ## Decision Rules (High Leverage) |
| 42 | |
| 43 | - Any high-stakes operation must include "challenge-and-response" against single-point human misread. |
| 44 | - If authority gradient rises and dissent drops, treat this as a safety hazard. |
| 45 | - If a parameter conversion is manual and repetitive, redesign or hard-check it. |
| 46 | - If procedure compliance and practical reality diverge, document and resolve as system debt. |
| 47 | |
| 48 | ## Safety-Critical Anti-Patterns |
| 49 | |
| 50 | - Punish-first response culture suppressing reporting. |
| 51 | - "Experienced crew means safe by default" assumption. |
| 52 | - Hidden mixed standards (units/procedures/phrasings) across tools and teams. |
| 53 | - Informal shortcuts becoming normal in marginal conditions. |
| 54 | |
| 55 | ## Cross-Domain Transfer Boundary |
| 56 | |
| 57 | - Human-factors insights transfer well to other high-risk domains (medicine, infra ops, incident response), |
| 58 | but direct policy transfer requires local regulatory and culture checks. |
| 59 | |
| 60 | ## Domain DoD v1 |
| 61 | |
| 62 | This domain is "done v1" when: |
| 63 | - playbook exists with TEM/CRM/ADM/JustCulture/LOSA contracts; |
| 64 | - reading map exists with first-party grounding; |
| 65 | - one curated rules set is available for operational reuse; |
| 66 | - status artifact and index routing are synchronized. |
| 67 | |
| 68 | |