Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Apply Granovetter's embeddedness theory to analyze how economic behavior is embedded in ongoing social relations, avoiding both over-socialized and under-socialized accounts. Use this skill when the user needs to explain why market transactions deviate from pure economic rationality, analyze how trust and social ties shape business dealings, evaluate structural vs relational embeddedness in inter-firm networks, or when they ask 'why do firms prefer existing partners over cheaper alternatives', '
.claude/skills/asgard-ai-platform-grad-embeddedness/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-04 | ✗→✓ | ▲ Improved | -28% | 0% |
| case-05 | ✗→✓ | ▲ Improved | -34% | 0% |
| case-19 | ✗→✓ | ▲ Improved | 35% | 0% |
| case-24 | ✗→✓ | ▲ Improved | -11% | 0% |
| case-01 | ✓→✓ | = Same ✓ | 44% | 0% |
Granovetter (1985) argued that economic action is embedded in concrete, ongoing systems of social relations — it is neither driven by atomized rational calculation (under-socialized view) nor by internalized cultural norms (over-socialized view). This "new economic sociology" reframes markets as social structures where trust, reputation, and network position shape transactions.
IRON LAW: Economic behavior is NEITHER purely rational NOR purely
socially determined — it is embedded in ongoing social relations. Any
analysis that treats actors as either atomized utility-maximizers or
cultural automatons violates the embeddedness thesis.Key assumptions:
Define the transaction, exchange, or economic behavior under analysis. Specify the actors and the market context.
| View | Assumption | Problem | |------|-----------|---------| | Under-socialized (neoclassical economics) | Actors are atomized, rational, self-interested | Ignores trust, reputation, ongoing relationships | | Over-socialized (Parsonian sociology) | Actors follow internalized norms automatically | Ignores agency, strategy, network variation | | Embeddedness (Granovetter) | Action is embedded in ongoing social relations | The middle ground — empirically trace the relationships |
| Dimension | Focus | Key Questions | |-----------|-------|---------------| | Structural embeddedness | Network architecture | How does the overall network topology (density, centrality, clustering) shape behavior? | | Relational embeddedness | Dyadic tie quality | How do trust, reciprocity, and history between specific pairs of actors affect transactions? |
Assess how embeddedness affects efficiency, opportunism, innovation, and lock-in.
markdown## Embeddedness Analysis: [Context] ### Economic Action - Transaction: [what is being exchanged] - Actors: [who is involved] - Market context: [industry, competitive structure] ### Socialization Assessment - Under-socialized explanation: [what pure economics would predict] - Over-socialized explanation: [what pure cultural determinism would predict] - Embeddedness explanation: [how social relations actually shape the behavior] ### Embeddedness Dimensions | Dimension | Evidence | Effect on Behavior | |-----------|----------|-------------------| | Structural embeddedness | [network position, density] | [how it constrains/enables] | | Relational embeddedness | [trust, history, reciprocity] | [how it constrains/enables] | ### Benefits and Costs of Embeddedness | Benefits | Costs | |----------|-------| | [trust reduces transaction costs] | [lock-in, obligation, insularity] | ### Implications 1. [How embeddedness explains the observed deviation from pure market logic] 2. [Risks of over-embeddedness or under-embeddedness]
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | pass→pass | 14,730 | 15,607 | +6% | 1 | 1 | 0% | 2,148 | 3,099 | +44% | 0 | 0 | — |
case-02 | pass→pass | 17,105 | 13,930 | -19% | 1 | 1 | 0% | 2,737 | 3,273 | +20% | 0 | 0 | — |
case-03 | pass→pass | 18,203 | 13,753 | -24% | 1 | 1 | 0% | 2,609 | 3,189 | +22% | 0 | 0 | — |
case-04 | fail→pass | 34,295 | 16,875 | -51% | 1 | 1 | 0% | 5,102 | 3,660 | -28% | 0 | 0 | — |
case-05 | fail→pass | 48,134 | 26,833 | -44% | 1 | 1 | 0% | 7,393 | 4,911 | -34% | 0 | 0 | — |
case-06 | pass→pass | 31,981 | 19,064 | -40% | 1 | 1 | 0% | 4,431 | 4,029 | -9% | 0 | 0 | — |
case-07 | pass→pass | 22,782 | 21,201 | -7% | 1 | 1 | 0% | 3,397 | 3,836 | +13% | 0 | 0 | — |
case-08 | fail→fail | 30,547 | 21,773 | -29% | 1 | 1 | 0% | 4,313 | 4,347 | +1% | 0 | 0 | — |
case-09 | pass→pass | 28,512 | 18,354 | -36% | 1 | 1 | 0% | 4,196 | 3,964 | -6% | 0 | 0 | — |
case-10 | pass→pass | 41,304 | 24,278 | -41% | 1 | 1 | 0% | 5,892 | 4,713 | -20% | 0 | 0 | — |
case-11 | pass→pass | 29,340 | 20,036 | -32% | 1 | 1 | 0% | 4,494 | 4,192 | -7% | 0 | 0 | — |
case-12 | pass→pass | 39,114 | 20,086 | -49% | 1 | 1 | 0% | 5,949 | 4,199 | -29% | 0 | 0 | — |
case-13 | pass→pass | 36,036 | 20,401 | -43% | 1 | 1 | 0% | 5,610 | 4,340 | -23% | 0 | 0 | — |
case-14 | pass→pass | 33,297 | 19,418 | -42% | 1 | 1 | 0% | 4,852 | 4,116 | -15% | 0 | 0 | — |
case-15 | pass→pass | 35,991 | 22,651 | -37% | 1 | 1 | 0% | 5,623 | 4,258 | -24% | 0 | 0 | — |
case-16 | pass→pass | 32,098 | 23,146 | -28% | 1 | 1 | 0% | 5,209 | 4,457 | -14% | 0 | 0 | — |
case-17 | pass→pass | 38,920 | 19,122 | -51% | 1 | 1 | 0% | 5,404 | 4,215 | -22% | 0 | 0 | — |
case-18 | pass→pass | 18,982 | 20,677 | +9% | 1 | 1 | 0% | 2,924 | 4,156 | +42% | 0 | 0 | — |
case-19 | fail→pass | 20,742 | 22,346 | +8% | 1 | 1 | 0% | 3,154 | 4,273 | +35% | 0 | 0 | — |
case-20 | pass→pass | 14,802 | 15,448 | +4% | 1 | 1 | 0% | 2,518 | 3,650 | +45% | 0 | 0 | — |
case-21 | pass→pass | 15,027 | 13,707 | -9% | 1 | 1 | 0% | 2,067 | 3,138 | +52% | 0 | 0 | — |
case-22 | pass→pass | 20,655 | 21,575 | +4% | 1 | 1 | 0% | 3,342 | 4,737 | +42% | 0 | 0 | — |
case-23 | pass→pass | 51,318 | 23,098 | -55% | 1 | 1 | 0% | 6,832 | 4,715 | -31% | 0 | 0 | — |
case-24 | fail→pass | 28,081 | 19,104 | -32% | 1 | 1 | 0% | 4,677 | 4,155 | -11% | 0 | 0 | — |
DecimalAI ran this skill against gemini-3.6-flash twice over the same eval suite — once with the skill loaded and once without — and compared the two runs case by case. 24 cases were attempted. The headline lift of +17 percentage points is the difference between those two pass rates over the 24 comparable cases.
Without the skill loaded, the model failed this case. With it loaded, the same prompt on the same model passed. This is one improved case from the latest verified run; every case, including any that regressed, is in the table above.
Other measured skills in the registry, with their headline benchmark lift.