Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Apply Weick's sensemaking theory to analyze how individuals and organizations construct meaning from ambiguous situations. Use this skill when the user needs to analyze organizational responses to crises, understand how interpretive frames shape action, diagnose breakdowns in collective understanding, or when they ask 'how did they interpret this situation', 'why did the organization fail to see the warning signs', or 'how do people make sense of disruption'.
.claude/skills/asgard-ai-platform-grad-sensemaking/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-01 | ✗→✓ | ▲ Improved | 26% | 0% |
| case-02 | ✗→✓ | ▲ Improved | 10% | 0% |
| case-03 | ✗→✓ | ▲ Improved | -18% | 0% |
| case-04 | ✗→✓ | ▲ Improved | 8% | 0% |
| case-07 | ✗→✓ | ▲ Improved | 21% | 0% |
Sensemaking theory, developed by Karl Weick, explains how people structure the unknown by placing stimuli into frameworks that enable comprehension and action. It is fundamentally retrospective, social, and ongoing — people discover what they think by looking at what they have done.
IRON LAW: Sensemaking is retrospective — people make sense of situations
AFTER they act, not before. Any analysis that assumes actors first
understood the situation and then decided what to do reverses the
causal order of sensemaking.Key assumptions:
Select the ambiguous or disruptive event that triggered sensemaking. Define the temporal boundaries.
| Property | Question to Ask | |----------|----------------| | Identity construction | How did actors' sense of "who we are" shape interpretation? | | Retrospect | What past actions were reinterpreted to make the situation intelligible? | | Enactment | How did actors' actions create the environment they then responded to? | | Social | How was meaning negotiated among group members? | | Ongoing | How did sensemaking evolve as the situation unfolded? | | Extracted cues | Which cues were noticed and which were ignored? Why? | | Plausibility over accuracy | What "good enough" story was adopted? What accuracy was sacrificed? |
Locate cosmology episodes (total loss of meaning), commitment traps, or situations where enactment created self-fulfilling prophecies.
Evaluate how the sense made (or not made) shaped organizational action and outcomes.
markdown## Sensemaking Analysis: [Context] ### Triggering Event - Event: [description of the ambiguous/disruptive situation] - Temporal scope: [when sensemaking began and stabilized] ### Seven Properties Assessment | Property | Evidence | Impact on Interpretation | |----------|----------|------------------------| | Identity construction | [evidence] | [how it shaped meaning] | | Retrospect | [evidence] | [how it shaped meaning] | | Enactment | [evidence] | [how it shaped meaning] | | Social | [evidence] | [how it shaped meaning] | | Ongoing | [evidence] | [how it shaped meaning] | | Extracted cues | [evidence] | [how it shaped meaning] | | Plausibility | [evidence] | [how it shaped meaning] | ### Sensemaking Breakdowns - [Where and why meaning collapsed or was distorted] ### Consequences 1. [How the sense made shaped subsequent action] 2. [What alternative interpretations were foreclosed]
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→pass | 18,169 | 19,293 | +6% | 1 | 1 | 0% | 2,865 | 3,609 | +26% | 0 | 0 | — |
case-02 | fail→pass | 26,632 | 22,056 | -17% | 1 | 1 | 0% | 4,059 | 4,455 | +10% | 0 | 0 | — |
case-03 | fail→pass | 29,560 | 17,867 | -40% | 1 | 1 | 0% | 4,238 | 3,475 | -18% | 0 | 0 | — |
case-04 | fail→pass | 19,672 | 17,914 | -9% | 1 | 1 | 0% | 3,251 | 3,523 | +8% | 0 | 0 | — |
case-05 | pass→pass | 11,575 | 11,642 | +1% | 1 | 1 | 0% | 1,955 | 3,053 | +56% | 0 | 0 | — |
case-06 | pass→fail | 13,641 | 14,561 | +7% | 1 | 1 | 0% | 2,861 | 3,095 | +8% | 0 | 0 | — |
case-07 | fail→pass | 22,740 | 19,086 | -16% | 1 | 1 | 0% | 3,165 | 3,832 | +21% | 0 | 0 | — |
case-08 | pass→pass | 23,807 | 24,028 | +1% | 1 | 1 | 0% | 2,861 | 4,878 | +70% | 0 | 0 | — |
case-09 | fail→pass | 31,334 | 20,719 | -34% | 1 | 1 | 0% | 4,679 | 4,105 | -12% | 0 | 0 | — |
case-10 | pass→pass | 18,168 | 15,416 | -15% | 1 | 1 | 0% | 2,724 | 3,321 | +22% | 0 | 0 | — |
case-11 | fail→pass | 17,778 | 16,030 | -10% | 1 | 1 | 0% | 2,647 | 3,524 | +33% | 0 | 0 | — |
case-12 | pass→pass | 18,737 | 20,752 | +11% | 1 | 1 | 0% | 2,770 | 4,096 | +48% | 0 | 0 | — |
case-13 | pass→pass | 20,269 | 15,292 | -25% | 1 | 1 | 0% | 2,945 | 3,303 | +12% | 0 | 0 | — |
case-14 | pass→pass | 17,228 | 20,211 | +17% | 1 | 1 | 0% | 2,700 | 4,119 | +53% | 0 | 0 | — |
case-15 | pass→pass | 26,474 | 15,392 | -42% | 1 | 1 | 0% | 3,624 | 3,195 | -12% | 0 | 0 | — |
case-16 | pass→pass | 15,711 | 17,029 | +8% | 1 | 1 | 0% | 2,497 | 3,451 | +38% | 0 | 0 | — |
case-17 | fail→pass | 16,032 | 14,266 | -11% | 1 | 1 | 0% | 2,651 | 3,170 | +20% | 0 | 0 | — |
case-18 | pass→pass | 25,966 | 29,563 | +14% | 1 | 1 | 0% | 3,651 | 5,675 | +55% | 0 | 0 | — |
case-19 | pass→pass | 19,203 | 17,705 | -8% | 1 | 1 | 0% | 2,880 | 3,458 | +20% | 0 | 0 | — |
case-20 | fail→pass | 12,945 | 15,204 | +17% | 1 | 1 | 0% | 1,904 | 3,145 | +65% | 0 | 0 | — |
case-21 | pass→pass | 16,262 | 17,732 | +9% | 1 | 1 | 0% | 2,285 | 3,465 | +52% | 0 | 0 | — |
case-22 | pass→pass | 21,377 | 18,469 | -14% | 1 | 1 | 0% | 2,787 | 3,585 | +29% | 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. 22 cases were attempted. The headline lift of +36 percentage points is the difference between those two pass rates over the 22 comparable cases. 1 case got worse with the skill loaded, and it is included in that figure.
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.