@human-avatar/skills-for-humanity 1.0.0

package/skills/resource/SKILL.md

@@ -0,0 +1,78 @@
+---
+name: resource
+description: "Entry point for the resource toolkit. Routes to the right resource skill based on your situation. Use when you say 'resource', 'capacity', 'bottleneck', 'allocation', 'where should we focus', 'where are we wasting', 'highest leverage', 'what's slowing us down', or want resource reasoning applied without knowing which specific tool fits."
+---
+
+# Resource
+
+Applies resource reasoning to allocation, constraints, leverage, and waste. Diagnoses what kind of resource work is needed and applies the right tool.
+
+## Which tool fits
+
+| You need to... | Tool |
+|---|---|
+| Distribute limited resources across competing needs | allocation-analysis |
+| Find what is actually constraining throughput | bottleneck-analysis |
+| Find the highest-leverage use of available resources | leverage-mapping |
+| Find where resources are being lost, duplicated, or underused | waste-audit |
+
+## Routing Decision
+
+- **Have limited resources and competing priorities — need to allocate** → allocation-analysis
+- **Things are moving slowly — need to find the actual constraint** → bottleneck-analysis
+- **Want to maximize impact with what you have** → leverage-mapping
+- **Sense that effort isn't producing commensurate output** → waste-audit
+- **Unclear** → bottleneck-analysis; finding the constraint usually determines how to allocate and where the leverage is
+
+## Confirm Direction
+
+After diagnosing which tool fits, present the recommendation before executing:
+
+> My read: **[diagnosed tool]** — one sentence on why it fits.
+
+- **A) Yes, run that tool**
+- **B) Show me all options** — list every skill in this category with one-line descriptions
+- **C) Quick version** — lighter-weight alternative for this situation, if one exists
+- **D) Re-diagnose** — describe the situation differently for a second read
+
+Wait for their selection before proceeding.
+
+---
+
+## Allocation Analysis
+
+*Distributes limited resources across competing needs.*
+
+Make the trade-offs explicit rather than implicit. Map all competing claims on the resource. For each: what is the expected return on investment? What is the cost of under-resourcing? Are any allocations producing diminishing returns? Are any starved below a minimum threshold for effectiveness? Allocation decisions made without explicit trade-off analysis tend to satisfy the loudest voice rather than the best use.
+
+**Output:** Allocation map with expected returns, identified diminishing returns, minimum thresholds, and the recommended distribution with explicit trade-off reasoning.
+
+---
+
+## Bottleneck Analysis
+
+*Identifies what is actually constraining throughput.*
+
+Apply Theory of Constraints logic: the system can only move as fast as its slowest point. Adding resources anywhere except the bottleneck doesn't improve throughput. Identify the current bottleneck: where does work queue up, where does it wait, where is output consistently below demand? Verify it's the actual bottleneck and not a symptom. Once identified: what would it take to elevate it?
+
+**Output:** Bottleneck identified with evidence, verification that it's the system constraint (not a symptom), and the options to elevate it with effort estimates.
+
+---
+
+## Leverage Mapping
+
+*Finds the highest-leverage use of available resources.*
+
+Leverage is where the same input produces the most output. Not all investments are equal — some create multiplying effects (better systems, better people, better tools) while others are purely linear. For each candidate use of resources: is the return linear (1 unit in → 1 unit out) or multiplying (1 unit in → many units out over time)? Prioritize multiplying investments, especially those that unblock other work.
+
+**Output:** Candidate uses of resources mapped by leverage type (linear vs. multiplying), with the highest-leverage options ranked and reasoning for each.
+
+---
+
+## Waste Audit
+
+*Finds where resources are being lost, duplicated, or underused.*
+
+Apply the seven wastes to knowledge work: (1) Overproduction — work done before it's needed, (2) Waiting — time spent blocked or idle, (3) Transport — handoffs that add no value, (4) Over-processing — more work than the task requires, (5) Inventory — work in progress that isn't being acted on, (6) Motion — effort that doesn't produce output, (7) Defects — rework from problems earlier in the process. Find the biggest waste category and its source.
+
+**Output:** Waste audit across all seven categories, the biggest waste sources, and the highest-impact reductions.

package/skills/resource-allocation-analysis/SKILL.md

@@ -0,0 +1,71 @@
+---
+name: resource-allocation-analysis
+description: "Distributes limited resources across competing needs — making the trade-offs explicit rather than implicit. Triggers: 'resource allocation', 'how do we distribute this', 'competing priorities', 'trade-off analysis', 'how do we split this', 'allocation decision'."
+---
+
+# Resource Allocation Analysis
+
+Every allocation is a trade-off — giving to one thing means giving less to another. The problem is that most allocations are made implicitly, leaving the trade-offs invisible. Making trade-offs explicit forces honest prioritisation and prevents the political habit of pretending everything can be fully funded.
+
+---
+
+## Your Process
+
+**Step 1: Inventory Available Resources**
+Name and quantify the resources to be allocated: budget, headcount, time, capacity. Be precise about what is actually available — not what is desired.
+
+**Step 2: List All Competing Claims**
+Every demand on the resource. Include maintenance and ongoing commitments, not just new initiatives. Claims that are implicitly assumed to be funded should be made explicit here.
+
+**Step 3: Assess Each Claim**
+For each claim: what is the strategic priority (how directly does this serve the most important goals)? What is the cost of under-resourcing it (what breaks, slows, or is lost if it receives less)?
+
+**Step 4: Identify Constraints**
+Are there minimums (must have at least X to function), maximums (more than Y produces no additional value), or dependencies (A must be funded before B makes sense)?
+
+**Step 5: Draft an Allocation**
+Distribute the available resource across claims. At this stage, make the trade-offs explicit: write down what each claim gives up under this draft allocation.
+
+**Step 6: Sense-Check Against Overall Goals**
+Does this allocation, taken as a whole, serve the most important outcomes? Where is the allocation driven by politics or inertia rather than strategic priority?
+
+---
+
+## Human Check-in
+
+Before proceeding, ask the user:
+
+**How do you want to run this?**
+
+- **A) Full analysis** — complete all steps, reasoning shown throughout
+- **B) Key findings only** — bottom-line output, skip step-by-step detail
+- **C) Trade-off table only** — what each allocation gives up, skip the full recommendation
+- **D) Refine the framing** — adjust what we're analyzing before starting
+
+Proceed based on their choice.
+
+## Output Format
+
+### Available Resources
+| Resource | Total Available |
+|----------|----------------|
+| ... | ... |
+
+### Competing Claims
+| Claim | Strategic Priority | Cost of Under-Resourcing | Constraints |
+|-------|------------------|-------------------------|-------------|
+| ... | High / Medium / Low | ... | Min / Max / Dependency |
+
+### Draft Allocation
+| Claim | Allocation | Trade-off (what it gives up) |
+|-------|-----------|------------------------------|
+| ... | ... | ... |
+
+### Strategic Alignment Check
+Does this allocation serve the most important outcomes? Where does it diverge from strategic priority — and is that divergence justified?
+
+---
+
+## Notes
+
+The most useful output is the trade-off column — if the trade-offs can't be written clearly, the allocation hasn't been thought through. Force each trade-off to be named before the allocation is finalised.

package/skills/resource-bottleneck-analysis/SKILL.md

@@ -0,0 +1,76 @@
+---
+name: resource-bottleneck-analysis
+description: "Identifies what is actually constraining throughput — using Theory of Constraints logic: the system can only move as fast as its slowest point. Triggers: 'bottleneck analysis', 'what's actually slowing this down', 'where's the constraint', 'theory of constraints', 'find the bottleneck', 'why are we moving slowly'."
+---
+
+# Bottleneck Analysis
+
+A system can only move as fast as its slowest stage. Improving any stage other than the bottleneck does not improve overall throughput — it just creates more work waiting in front of the constraint. Finding and addressing the bottleneck is the highest-leverage intervention available.
+
+---
+
+## Your Process
+
+**Step 1: Map the Process or Value Chain**
+Describe all the stages in sequence. What enters each stage, what happens, and what exits? The map should cover the full flow from input to output.
+
+**Step 2: Identify Throughput per Stage**
+For each stage: what is the throughput rate? How much work can it process per unit of time? Where does work visibly queue up, wait, or accumulate? Queuing upstream of a stage is the most reliable indicator of a bottleneck.
+
+**Step 3: Identify the Current Bottleneck**
+The bottleneck is the stage with the lowest throughput that gates everything downstream. Name it specifically. Everything else in the system is constrained by its capacity.
+
+**Step 4: Exploit the Bottleneck (No New Investment)**
+Before investing in more capacity: are we getting maximum output from the bottleneck as it currently exists? What work is arriving at the bottleneck that shouldn't be there? Can quality gates be moved earlier to protect bottleneck time? Can idle time at the bottleneck be eliminated?
+
+**Step 5: Subordinate Everything Else to the Bottleneck**
+Non-bottleneck stages should operate at the rate the bottleneck can consume. Optimising non-bottleneck stages creates inventory, not throughput. Identify any optimisations elsewhere in the system that are making the bottleneck problem worse.
+
+**Step 6: Elevate the Bottleneck**
+Only after Steps 4–5: what investment would increase the bottleneck's capacity? What would it cost, and what throughput improvement would result?
+
+---
+
+## Human Check-in
+
+Before proceeding, ask the user:
+
+**How do you want to run this?**
+
+- **A) Full analysis** — complete all steps, reasoning shown throughout
+- **B) Key findings only** — bottom-line output, skip step-by-step detail
+- **C) Constraint identification only** — what's actually limiting throughput, skip the relief analysis
+- **D) Refine the framing** — adjust what we're analyzing before starting
+
+Proceed based on their choice.
+
+## Output Format
+
+### Process Map
+[Stage 1] → [Stage 2] → [Stage 3] → ... → [Output]
+
+### Throughput per Stage
+| Stage | Throughput Rate | Queue Size / Wait Time | Bottleneck? |
+|-------|----------------|----------------------|-------------|
+| ... | ... | ... | Yes / No |
+
+### Current Bottleneck
+**Stage:** [Name]
+**Why it's the constraint:** [Throughput evidence]
+
+### Exploit Options (No New Investment Required)
+- [Specific action to get more from the existing bottleneck]
+
+### Subordination Recommendations
+- [What should stop being optimised because it feeds the bottleneck]
+
+### Elevation Options (Investment Required)
+| Option | Investment | Throughput Improvement |
+|--------|-----------|----------------------|
+| ... | ... | ... |
+
+---
+
+## Notes
+
+The bottleneck moves after you fix it — the next constraint becomes the new bottleneck. Repeat the analysis after each intervention rather than assuming the system is now optimised.

package/skills/resource-leverage-mapping/SKILL.md

@@ -0,0 +1,69 @@
+---
+name: resource-leverage-mapping
+description: "Finds the highest-leverage use of available resources — where the same input produces the most output. Triggers: 'resource leverage', 'highest-leverage use', 'where should we put our energy', 'maximize impact', 'leverage mapping'."
+---
+
+# Resource Leverage Mapping
+
+Not all uses of a resource are equal. Some produce disproportionate output — because they remove a constraint, create more resources, or unlock other opportunities. Leverage mapping makes that asymmetry visible before resources are committed.
+
+---
+
+## Your Process
+
+**Step 1: Inventory Available Resources**
+List all meaningful resources available: time, money, people, attention, relationships, existing assets, reputation. Be specific — "the engineering team" is less useful than "three senior engineers with 20% slack capacity."
+
+**Step 2: List All Candidate Uses**
+For each resource, what are all the plausible ways it could be deployed? Do not filter yet — generate a full list of options.
+
+**Step 3: Estimate Output per Unit of Input**
+For each candidate use: what is the likely output for a given unit of input? This doesn't need to be precise — a rough order-of-magnitude estimate is sufficient. The goal is to find the outliers.
+
+**Step 4: Identify Multiplier Effects**
+Which uses of a resource create more resources, unlock additional capacity, or enable other uses? These are the highest-leverage options. Examples: building a relationship that opens a distribution channel; shipping a feature that funds the next two.
+
+**Step 5: Find Underused High-Leverage Resources**
+Which available resources are currently underused relative to their potential leverage? Relationships, existing data, attention from a key person, and existing assets are commonly overlooked.
+
+**Step 6: Recommend the Highest-Leverage Allocation**
+Given the analysis, what is the best deployment of the available resources?
+
+---
+
+## Human Check-in
+
+Before proceeding, ask the user:
+
+**How do you want to run this?**
+
+- **A) Full analysis** — complete all steps, reasoning shown throughout
+- **B) Key findings only** — bottom-line output, skip step-by-step detail
+- **C) Highest leverage use only** — single best allocation of the scarcest resource
+- **D) Refine the framing** — adjust what we're analyzing before starting
+
+Proceed based on their choice.
+
+## Output Format
+
+### Resource Inventory
+| Resource | Available Quantity / Capacity |
+|----------|------------------------------|
+| ... | ... |
+
+### Candidate Uses with Output Estimates
+| Resource | Use | Estimated Output per Unit Input | Multiplier Effect? |
+|----------|-----|--------------------------------|--------------------|
+| ... | ... | Low / Medium / High | Yes / No — [describe] |
+
+### Underused High-Leverage Resources
+- [Resource] — current use vs. potential leverage.
+
+### Recommended Allocation
+State the highest-leverage deployment with rationale. Make the trade-offs explicit — what is being deprioritised and why.
+
+---
+
+## Notes
+
+Run this before a planning cycle or before committing significant resources to a course of action. The most common finding is that a relationship or an existing asset is being underused relative to its potential leverage.

package/skills/resource-waste-audit/SKILL.md

@@ -0,0 +1,80 @@
+---
+name: resource-waste-audit
+description: "Finds where resources are being lost, duplicated, or underused — the seven wastes applied to knowledge work. Triggers: 'waste audit', 'where are we wasting resources', 'inefficiency audit', 'find the waste', 'what's being duplicated', 'resource leakage'."
+---
+
+# Resource Waste Audit
+
+In manufacturing, Toyota identified seven categories of waste. They apply equally to knowledge work — with a different surface appearance but the same underlying structure. The goal is to find where resources are consumed without producing value.
+
+---
+
+## The Seven Wastes in Knowledge Work
+
+| Waste | Description |
+|-------|-------------|
+| **Waiting** | Idle time between steps — work blocked pending a decision, review, or dependency |
+| **Overproduction** | Producing more than is needed — reports no one reads, features no one uses |
+| **Rework** | Fixing what should have been right — re-doing work due to unclear requirements or poor handoffs |
+| **Duplication** | The same work being done in two places — parallel efforts, re-discovered knowledge |
+| **Motion** | Unnecessary switching or handoffs — context switching, excessive meetings, process overhead |
+| **Inventory** | Work in progress that is not flowing — large backlogs, features built but not shipped |
+| **Over-processing** | More effort than the task requires — over-engineered solutions, unnecessary polish |
+
+---
+
+## Your Process
+
+**Step 1: Map the Workflow or Resource Allocation**
+Describe how work moves through the system — the full path from input to value delivery.
+
+**Step 2: Scan Each Waste Category**
+For each of the seven wastes: where does it appear in the workflow? Look for specific instances, not general impressions.
+
+**Step 3: Quantify Each Waste**
+Estimate roughly how much resource each waste consumes — time per week, headcount, or percentage of capacity. Rough estimates are fine; the goal is to rank, not to audit precisely.
+
+**Step 4: Root Cause per Waste**
+Why does this waste exist? Process design? Incentive structure? Unclear ownership? Identifying the root cause determines whether the fix is simple or systemic.
+
+**Step 5: Rank by Impact and Recommend**
+Which waste removal would free the most resource? Prioritise the top three and propose specific actions.
+
+---
+
+## Human Check-in
+
+Before proceeding, ask the user:
+
+**How do you want to run this?**
+
+- **A) Full analysis** — complete all steps, reasoning shown throughout
+- **B) Key findings only** — bottom-line output, skip step-by-step detail
+- **C) Biggest waste items only** — top 3 losses by magnitude, skip smaller inefficiencies
+- **D) Refine the framing** — adjust what we're analyzing before starting
+
+Proceed based on their choice.
+
+## Output Format
+
+### Waste Inventory
+| Waste Type | Where It Appears | Estimated Resource Cost | Root Cause |
+|-----------|-----------------|------------------------|------------|
+| Waiting | ... | ... | ... |
+| Overproduction | ... | ... | ... |
+| Rework | ... | ... | ... |
+| Duplication | ... | ... | ... |
+| Motion | ... | ... | ... |
+| Inventory | ... | ... | ... |
+| Over-processing | ... | ... | ... |
+
+### Top 3 Waste Reductions (Ranked by Impact)
+1. **[Waste type]** — [Specific action] — [Expected resource freed]
+2. **[Waste type]** — [Specific action] — [Expected resource freed]
+3. **[Waste type]** — [Specific action] — [Expected resource freed]
+
+---
+
+## Notes
+
+Rework and duplication are usually the most expensive wastes in knowledge work, but waiting is often the most demoralising. Address the highest-cost waste first, but don't ignore the one most affecting morale.

package/skills/sensory/SKILL.md

@@ -0,0 +1,68 @@
+---
+name: sensory
+description: "Entry point for the sensory observation toolkit. Routes to the right observational skill based on your situation. Use when you say 'sensory', 'what are we missing', 'look more carefully', 'separate signal from noise', 'too much information', 'what actually matters here', 'observe this carefully', or want observational precision applied without knowing which specific tool fits."
+---
+
+# Sensory
+
+Applies disciplined observation and attention to any situation. Diagnoses what kind of observational work is needed and applies the right tool.
+
+## Which tool fits
+
+| You need to... | Tool |
+|---|---|
+| Find specific details that are being overlooked | detail-mining |
+| Separate what actually matters from background noise | signal-detection |
+| Apply disciplined observation before interpreting | structured-observation |
+
+## Routing Decision
+
+- **Have a situation and suspect important details are being missed** → detail-mining
+- **Overwhelmed with information and need to find what matters** → signal-detection
+- **Rushing to interpret before fully observing** → structured-observation
+- **Unclear** → structured-observation first; disciplined observation before mining or filtering
+
+---
+
+## Confirm Direction
+
+After diagnosing which tool fits, present the recommendation before executing:
+
+> My read: **[diagnosed tool]** — one sentence on why it fits.
+
+- **A) Yes, run that tool**
+- **B) Show me all options** — list every skill in this category with one-line descriptions
+- **C) Quick version** — lighter-weight alternative for this situation, if one exists
+- **D) Re-diagnose** — describe the situation differently for a second read
+
+Wait for their selection before proceeding.
+
+---
+
+## Detail Mining
+
+*Finds specific details being overlooked.*
+
+Slow down. The most important information is often present but not being registered. Go back to primary sources — the actual data, the actual words, the actual behavior. For each element: what specific, concrete, observable detail is present? What exactly is being said, not what it means? Generate a dense inventory of specifics before drawing conclusions. Vague observations produce vague conclusions.
+
+**Output:** Dense inventory of specific, concrete details. The details most likely to be overlooked and their implications.
+
+---
+
+## Signal Detection
+
+*Separates meaningful signal from background noise.*
+
+Not all information is equally relevant. Classify each piece of information: (1) Strong signal — directly relevant to the question, (2) Weak signal — potentially relevant, worth watching, (3) Noise — present but irrelevant. For each strong signal: what does it indicate? For each weak signal: what would make it stronger or weaker? The goal is attention management — what should be acted on, watched, and ignored.
+
+**Output:** Signal/noise classification for all inputs. Strong signals and their implications. Weak signals to monitor.
+
+---
+
+## Structured Observation
+
+*Applies disciplined observation by suspending interpretation.*
+
+Most people observe and interpret simultaneously — which means interpretation shapes what they see. Suspend interpretation: for this exercise, describe only what you observe, not what it means. What is actually present? What is actually happening? What are the exact words, behaviors, or outputs? After observation is complete, interpretation can begin — but it starts from a richer, more accurate base.
+
+**Output:** Pure observation record (no interpretation). Then: interpretation layer applied to the observation record, now clearly separated from what was actually seen.

package/skills/sensory-detail-mining/SKILL.md

@@ -0,0 +1,70 @@
+---
+name: sensory-detail-mining
+description: "Finds specific details being overlooked — the most important information is often present but not being registered. Triggers: 'what are we missing', 'go deeper on this', 'find the details', 'be more specific', 'what exactly is happening', 'ground this in specifics'."
+---
+
+# Detail Mining
+
+Abstractions are useful — but they lose the specific detail that often contains the real insight. "Users are frustrated" is an abstraction that conceals which users, in which moment, doing what, saying what exactly. Detail mining forces that concealment back into the open.
+
+---
+
+## Your Process
+
+**Step 1: Take the Current Description**
+Work with whatever account, analysis, or summary exists. This is the starting material — it contains the abstractions to excavate.
+
+**Step 2: Identify Where It's Abstract**
+Mark every place the description uses categories, summaries, or generalisations instead of specific observed instances. Words like "often," "users," "usually," "issues," "problems," and "feedback" are abstraction signals.
+
+**Step 3: Force Specificity on Each Abstraction**
+For each abstraction: what are the actual, specific instances behind it? Name them. Quote them if possible. Specify who, when, what exactly.
+- Instead of: "Users are frustrated."
+- Write: "3 users in session recordings said 'I don't understand this button' and clicked it twice before abandoning the flow."
+
+**Step 4: Recover Ignored Background Details**
+What is present in the situation but not described — treated as taken-for-granted background? List these. They are often invisible because everyone assumes they are known.
+
+**Step 5: Surface Absences as Details**
+What should be there but isn't? Absence is a detail. A missing response, a skipped step, a field left blank — these are observations, not gaps in data.
+
+**Step 6: Read What the Specifics Reveal**
+Now that you have the specifics: what do they show that the abstractions concealed? What changes about your understanding?
+
+---
+
+## Human Check-in
+
+Before proceeding, ask the user:
+
+**How do you want to run this?**
+
+- **A) Full analysis** — complete all steps, reasoning shown throughout
+- **B) Key findings only** — bottom-line output, skip step-by-step detail
+- **C) One overlooked detail** — the single most important thing being missed
+- **D) Refine the framing** — adjust what we're analyzing before starting
+
+Proceed based on their choice.
+
+## Output Format
+
+### Abstractions Identified
+| Abstraction | Specific Instances Behind It |
+|-------------|------------------------------|
+| "users are frustrated" | [exact quote / behaviour / timestamp] |
+| ... | ... |
+
+### Ignored Background Details
+- [Detail treated as given, not described]
+
+### Notable Absences
+- [What should be present but isn't]
+
+### What the Specifics Reveal
+Paragraph summary: how does the specific picture differ from the abstract one? What new questions or insights emerge?
+
+---
+
+## Notes
+
+This skill is most useful immediately before a decision, a diagnosis, or a design choice — the moment when abstractions are about to drive action. Getting specific at that point can prevent expensive errors.

package/skills/sensory-signal-detection/SKILL.md

@@ -0,0 +1,68 @@
+---
+name: sensory-signal-detection
+description: "Separates meaningful signal from background noise — finding what actually matters among everything present. Triggers: 'what actually matters here', 'separate signal from noise', 'too much information', 'find the signal', 'what should I focus on', 'what's relevant'."
+---
+
+# Signal Detection
+
+In any rich environment — data, feedback, conversation, a market — most of what is present is noise. Signal is what varies with the thing you're trying to understand; noise varies independently. The challenge is not finding more information, it's knowing which information is doing real work.
+
+---
+
+## Your Process
+
+**Step 1: Inventory Everything Present**
+List all the data, observations, or inputs available. Don't filter yet — complete the inventory first.
+
+**Step 2: Variance Test**
+For each item: does it vary with the outcome or phenomenon you're trying to understand? Signal co-varies with what you care about. Noise varies on its own schedule.
+
+**Step 3: Persistence Test**
+Is this item consistently present across time and contexts, or did it appear once? Persistent patterns are more likely to be signal. One-off observations may be noise, anomaly, or coincidence.
+
+**Step 4: Specificity Test**
+Is this item unique to this situation, or is it always present? Always-present background conditions are usually noise. What is specific to the case is more likely signal.
+
+**Step 5: Counterfactual Test**
+If this item changed or disappeared, would the outcome change? If yes: probable signal. If the outcome would be the same regardless: probable noise.
+
+**Step 6: Classify and Summarise**
+Assign a classification to each item and identify the top signals to act on.
+
+---
+
+## Human Check-in
+
+Before proceeding, ask the user:
+
+**How do you want to run this?**
+
+- **A) Full analysis** — complete all steps, reasoning shown throughout
+- **B) Key findings only** — bottom-line output, skip step-by-step detail
+- **C) Noise sources only** — what's obscuring the real signal
+- **D) Refine the framing** — adjust what we're analyzing before starting
+
+Proceed based on their choice.
+
+## Output Format
+
+### Element Inventory and Classification
+| Element | Varies with Outcome? | Persistent? | Specific? | Counterfactual? | Classification |
+|---------|---------------------|-------------|-----------|-----------------|----------------|
+| ... | ... | ... | ... | ... | Signal / Noise / Unclear |
+
+**Classifications:** Clear Signal / Probable Signal / Unclear / Probable Noise / Clear Noise
+
+### Top 3 Signals to Focus On
+1. [Signal] — rationale for prioritisation.
+2. [Signal] — rationale.
+3. [Signal] — rationale.
+
+### Notable Noise to Stop Tracking
+- Elements consuming attention without signal value.
+
+---
+
+## Notes
+
+When in doubt, classify as "unclear" rather than forcing a label — the act of flagging uncertainty is itself useful. Run this when a situation feels overwhelming or when a team is arguing about what matters.

package/skills/sensory-structured-observation/SKILL.md

@@ -0,0 +1,73 @@
+---
+name: sensory-structured-observation
+description: "Applies disciplined observation to a situation — suspending interpretation to see what's actually there before deciding what it means. Triggers: 'observe this carefully', 'structured observation', 'what do you actually see', 'suspend interpretation', 'look more carefully'."
+---
+
+# Structured Observation
+
+Most observation is interpretation in disguise. We perceive a situation and instantly explain it — but the explanation overwrites the raw data. Structured observation forces a separation between what can be directly seen and what we conclude from it.
+
+---
+
+## Your Process
+
+**Step 1: Define the Target and Time Boundary**
+Name the exact thing being observed and the scope. What counts as inside this observation, and what is out of scope?
+
+**Step 2: Separate Observation from Interpretation**
+Write only what can be directly observed — no inferences, no attributions of intent or cause. "User clicked back immediately" not "user was confused." Flag every sentence that is actually an inference and set it aside.
+
+**Step 3: Observe at Three Levels**
+- **Events** — what is happening? Discrete, specific occurrences.
+- **Patterns** — how is it happening? Recurring structure across events.
+- **Absences** — what is not happening that might be expected?
+
+**Step 4: Flag Surprising or Incongruent Observations**
+What doesn't fit? Where does something contradict expectations? These are the most information-rich observations — prioritise them.
+
+**Step 5: Generate Interpretations**
+Only after completing Steps 2–4: generate multiple possible interpretations for each key observation. Aim for at least two competing explanations.
+
+**Step 6: Identify the Most Testable Interpretation**
+Which interpretation makes the most specific, falsifiable prediction? That is the one to act on first.
+
+---
+
+## Human Check-in
+
+Before proceeding, ask the user:
+
+**How do you want to run this?**
+
+- **A) Full analysis** — complete all steps, reasoning shown throughout
+- **B) Key findings only** — bottom-line output, skip step-by-step detail
+- **C) Initial observations only** — what's actually there before any interpretation is applied
+- **D) Refine the framing** — adjust what we're analyzing before starting
+
+Proceed based on their choice.
+
+## Output Format
+
+### Observations
+| Level | Observation (no interpretation) |
+|-------|----------------------------------|
+| Event | ... |
+| Pattern | ... |
+| Absence | ... |
+
+### Surprising or Incongruent Observations
+- List each, with a note on why it is surprising.
+
+### Interpretations per Key Observation
+| Observation | Interpretation A | Interpretation B |
+|-------------|-----------------|-----------------|
+| ... | ... | ... |
+
+### Most Testable Interpretation
+State it as a prediction: "If [interpretation] is correct, then [specific observable consequence]."
+
+---
+
+## Notes
+
+Run this before diagnosis, analysis, or decision. The discipline has most value when you feel you already understand the situation — that feeling is usually a sign that interpretation has already overtaken observation.